<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Aleksi Aaltonen on Medium]]></title>
        <description><![CDATA[Stories by Aleksi Aaltonen on Medium]]></description>
        <link>https://medium.com/@aleksiaaltonen?source=rss-394d985f3bbf------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*etY_2KnfHfRbtJtRGJk4Mg.jpeg</url>
            <title>Stories by Aleksi Aaltonen on Medium</title>
            <link>https://medium.com/@aleksiaaltonen?source=rss-394d985f3bbf------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Tue, 17 Nov 2026 16:58:23 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@aleksiaaltonen/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Building an Academic Community — Principles, Practices, Problems]]></title>
            <link>https://aleksiaaltonen.medium.com/building-an-academic-community-principles-practices-problems-19246b3e70bb?source=rss-394d985f3bbf------2</link>
            <guid isPermaLink="false">https://medium.com/p/19246b3e70bb</guid>
            <category><![CDATA[community]]></category>
            <category><![CDATA[knowledge]]></category>
            <category><![CDATA[academia]]></category>
            <dc:creator><![CDATA[Aleksi Aaltonen]]></dc:creator>
            <pubDate>Mon, 16 Mar 2026 20:29:43 GMT</pubDate>
            <atom:updated>2026-03-16T22:30:52.985Z</atom:updated>
            <content:encoded><![CDATA[<h3><strong>Building an Academic Community — Principles, Practices, Problems</strong></h3><p><strong>Why do we need academic communities? I hold a view that knowledge is not stored in the papers we write but instead knowing happens when we make sense of the world and its phenomena in practice. This cannot happen without the ‘knowers’, that is, a community of scholars who read, debate, and build on each other’s work.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*reXqjdyIEv896a573tAvLg.png" /></figure><p>One could perhaps try reading, debating, and building on one’s own work only, yet historically building academic knowledge has been an inherently collective effort. In other word, academic knowledge lives in a scholarly community: in who talks to whom, what questions are asked, and which bodies of work travel across communities and outside academia.</p><p>Over the past two years, I’ve been involved in community building around the study of data in the field of information systems and, more generally, management disciplines. The efforts have crystallized into the <a href="http://datastudiesbibliography.org/">datastudiesbibliography.org</a> website, a mailing list, and a series of events that I and others have organized. At some point, a provisional label ‘data studies’ appeared to identify the emerging community.</p><p>In this post, I offer reflections from my experience on supporting academic community building, neither as a success story nor as a failure, but as work in progress. If you are involved in similar efforts, you may find them useful. We were also featured with my colleague <a href="https://www.linkedin.com/in/mstelmaszak/">Marta Stelmaszak</a> in the <a href="https://www.routledge.com/Social-Media-for-Research-Impact-How-Scholars-Can-Share-Ideas-Build-Networks-and-Make-a-Difference/Bogers-Young/p/book/9781032964072"><em>Social Media for Research Impact</em></a> book by <a href="https://www.linkedin.com/in/bogers/">Marcel Bogers</a> and <a href="https://www.linkedin.com/in/mkeyoung/">Mike Young</a> who have kindly allowed us to reproduce the box on data studies below (the book is a great resource for how to use social media for scholarly impact).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*AJI-Vd1pQZyxwOAHPL7jfA.jpeg" /><figcaption>Page 129 from<em> the Social Media for Research Impact</em> book by Marcel Bogers and Mike Young.</figcaption></figure><h3><strong>Three ‘Principles’ of Academic Community Building</strong></h3><p>The first principle of building <strong>an academic community is that you cannot build an academic community</strong>. A community is not a thing that can be built by putting bricks on top of each other, to create a durable structure such as a building. A community can only happen. But, <em>for a community to happen someone needs to create an infrastructure and draw people to it so that the community can happen</em>. This infrastructure can consist of many different things that people find useful for connecting with each other.</p><p>The second principle is that <strong>an academic community is not about me, you, or any other individual</strong>. I love to talk about my own research because it is the most important thing to me so why not try to make everyone else to talk about it, too? Well, turns out that <a href="https://medium.com/@aleksiaaltonen/i-do-better-research-than-you-on-respect-and-disagreement-in-academia-dadfb57b7302">they probably think the same about their own work</a> so whose work we are going to talk about in the end. <em>Academic community must be about our collective work.</em> Therefore, supporting community building is first and foremost about everybody else’s work and helping them to make their work visible — whether I personally agree with every detail of their work or not.</p><p>The third principle is that <strong>an academic community cannot happen unless it already exists</strong>. Rather than trying to give birth to a new community, I think it is better to <em>think academic community building as an effort to help a latent community to become visible to itself</em>. Our colleagues are not going to change their research interests just because we organize an event or set up a website. However, these and other similar efforts may draw the attention of those who share similar interests to a common platform so that they can advance their own academic endeavors as a community.</p><p>These principles can help identify opportunities to foster community building, design community-building interventions, and, importantly, to improvise as things start to evolve. In the following, I discuss a few lessons learned from what happens when things start to take off.</p><h3><strong>Potential Problems and How to Handle Them</strong></h3><p>First, setting up things is often easier than maintaining them. Therefore, it is important to design interventions so that you can maintain them without exhausting yourself. This is much easier if you get value out of the efforts for yourself, too. For instance, we first created a curated bibliography on the study of data with Marta to make sense of the literature for ourselves and then started posting about bibliography entries in social media (making and keeping the effort visible is a critical part of community building).</p><p>Second, whatever you build for the community should serve a practical purpose. While many things hopefully work, others may not be picked by your colleagues. It is important not to be afraid to kill off things that are just a drag and do not add value to your efforts. In the age of social media and shortening attention spans, people will forget failed experiments even before you have dared to killed them. Ultimately, even the whole community may dissolve as scholarship moves on. To me, this is fine. The purpose of an academic community is to support the creation of scholarly knowledge, not to sustain itself as an institution.</p><p>Third, mistakes will be made. You may occasionally irritate a colleague who thinks differently from you. The best way to avoid making mistakes is, of course, not to do anything. In entrepreneurship theory, this is called ‘failure without learning’. Instead, my approach is to seek reconciliation if conflicts happen, while not letting a few potential tensions to stop me from doing what I think is needed for a community to happen.</p><p>Finally, let me make a few remarks about what has worked for us.</p><h3><strong>Practices of Academic Community Building</strong></h3><p>First, it is difficult to do alone. You will probably need a few trusted friends to work together as a team that shares roughly the same vision. This is not just about having people to share the workload — although supporting community building is a lot of work (something that people who have not tried it are sometimes oblivious to). You may need skills that you are unlikely to possess alone, but, even more importantly, having someone to develop and evaluate potential interventions together is invaluable. Otherwise you easily succumb to self-doubts and get nothing done, or risk doing something that leads the whole effort astray.</p><p>Second, as the effort gains momentum, more people will want to have a say where the effort should go. This is where things can easily go wrong. An approach that has worked so far for me is to invite those who share a reasonably similar or compatible vision and are ready to put in real effort, to gradually get more and more say over the effort. This does not mean that you should not listen to people who just share their thoughts — yet, it is a mistake to try to please everyone’s whims. Only those who are hands-on involved in doing things usually understand what it takes to get things done.</p><p>Third, at some point, the community needs a succession plan — while some, often limited number of people need to do the heavy lifting in the beginning, it is healthy and practically important that the community finds a mechanism to hand over the responsibility to the next generation. For instance, special interest groups supported by academic associations can provide a suitable structure for this purpose.</p><h3><strong>Some Concluding Thoughts</strong></h3><p>If knowledge lives in academic communities rather than as stored in individual papers, then community building is not just a peripheral activity but a central part of how academic fields sustain themselves and keep producing relevant knowledge on new questions. I hope these remarks encourage others to notice latent communities around them and to explore ways of making those visible to themselves.</p><p><em>Thank you </em><a href="https://www.linkedin.com/in/oliviabenfeldt/"><em>Olivia Benfeldt</em></a><em> and Marta Stelmaszak for your thoughtful comments on the draft versions of this post.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=19246b3e70bb" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[A Gentle Guide to Programming Concepts for First-time Python Learners]]></title>
            <link>https://aleksiaaltonen.medium.com/a-gentle-guide-to-programming-concepts-for-first-time-python-learners-8e078dcdb449?source=rss-394d985f3bbf------2</link>
            <guid isPermaLink="false">https://medium.com/p/8e078dcdb449</guid>
            <category><![CDATA[learning-python]]></category>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[python]]></category>
            <dc:creator><![CDATA[Aleksi Aaltonen]]></dc:creator>
            <pubDate>Fri, 02 Jan 2026 22:47:42 GMT</pubDate>
            <atom:updated>2026-01-02T22:47:42.503Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*8LAb10fMYzA0wFNmyrC6BA.png" /></figure><p><strong>Why learning your first programming language is so difficult? It is difficult because you are learning not just one but two languages!</strong></p><p>In addition to whatever programming language you have chosen (or has been chosen) as your first programming language, you also need to learn to talk about programming languages. For instance, you may think that you are learning <strong>Python</strong> but in reality you are also learning a <strong>metalanguage</strong> about Python. This is because…</p><blockquote>…to learn programming, it is not enough to start thinking in a new, oddly formal language, but also to work with concepts such as ‘variable’, ‘expression’, ‘operator’, and many others that have a specific meaning in the context of programming.</blockquote><p>Many new concepts that you need to learn do not belong to any particular programming language but, instead, they are about programming and programming languages in general, that is, ‘in the abstract’. Your instructor and materials will use these new concepts to communicate ideas about the actual programming language that you are learning.</p><p>I wrote this brief guide to my MIS110 students at Stevens Institute of Technology, to make learning the metalanguage of programming easier for first-time Python programmers.</p><h4>Scripts and modules</h4><p>A Python <em>script</em> is a file that contains Python program code and can be executed by a Python interpreter. A Python script should be named with <em>.py</em> file extension at the end. Scripts are closely related to <em>modules</em> that are scripts that are written so that their parts (functions, classes, variables) can be used by other scripts or modules.</p><h4><strong>Statements</strong></h4><p>A Python script is made of statements. A <em>statement</em> is roughly equivalent to a line of code, although a statement can in some cases span several lines. Python statements are mostly made of three types of elements:</p><ul><li>Assignments</li><li>Expressions</li><li>Control statements and structures</li></ul><p>All these can include variables, literal values, operators, and function calls.</p><h4><strong><em>Assignments</em></strong></h4><p>An <em>assignment</em> assigns a value to a variable<em>. </em>A <em>variable</em> is a named entity that identifies data. For instance, the following assignment statement assigns the numerical value of <em>1</em> to the variable named <em>x:</em></p><pre>x = 1</pre><p>The numerical value (1) in the above statement is expressed as a <em>literal value </em>(or just <em>literal</em>), that is, a value that is directly written into the code. Literals can also be made of text but then they need to be enclosed in quotation marks to avoid confusing them with named entities such as variables in your code. There is no risk of confusing numerical literals with other entities because, in Python code, an entity name cannot start with a number. For instance, the following attempt to assign a string literal to a variable named <em>1st_author</em> will results in an error.</p><pre>1st_author = “Aaltonen, A.”</pre><p>Finally, the equal to sign <em>= </em>in the example is an assignment operator. An <em>operator</em> is a symbol that tells Python how to relate, such as compare or assign, different parts of a statement or expression to each other. In addition to the assignment operator, some of the most common types of operators in Python code are arithmetic, comparison, and logical operators.</p><h4><strong>Expressions</strong></h4><p>An <em>expression</em> is a part of code that produces a value. For instance, arithmetic expressions evaluate to a value, function calls return a value, and literals represent a value. In the following example, the expression <em>x + 1</em> evaluates to whatever number is stored in the variable <em>x</em> added by <em>1</em>.</p><pre>x + 1</pre><p>Assuming that <em>x </em>equals to <em>1</em> from the example above, the expression evaluates to 2, that is, you can mentally replace the expression <em>x + 1</em> with the number 2. We can combine the expression with an assignment in the same statement, to ‘store’ the value in the variable <em>y</em>.</p><pre>y = x + 1</pre><p>As I have hinted above, function calls can work as expressions (although they do other things, too), because they return a value. A <em>function call</em> executes another piece of code (function) that can take input parameter values and return output values. A <em>function</em> is a named block of code that processes the inputs and returns the outputs. For instance, the following call to a function named <em>chr() </em>returns a character ‘a’.</p><pre>chr(97)</pre><p><em>97</em> is a literal input value to the function—we could replace it with any other expression that evaluates to an appropriate numerical value. In other words, a solitary variable or literal is also an expression, because they represent a value.</p><pre>x  # The x is an expression.<br><br>&quot;This is an expression.&quot;</pre><p>If a function is associated with a specific object, we call it a <em>method</em> of that object. This is because Python is an object-oriented programming language. Even if you would not write your code in object-oriented way, you cannot avoid using methods in addition to functions in your code. For instance, to make a string all uppercase, you use it’s associated <em>upper() </em>method using the <em>dot notation</em>, that is,<em> </em>instead of passing the string as a parameter to the function, you add the function call to end of the object separated by ‘.’)<em>.</em></p><pre>&gt;&gt;&gt; &quot;This is a string object&quot;.upper()<br>&#39;THIS IS A STRING OBJECT&#39;</pre><p>Expressions can evaluate to many different types of values. These are known as data types<em>. A data type </em>is classification of what kind of value a variable can hold. Among different data types that an expression may produce, Boolean expressions evaluate to <em>True</em> or <em>False</em>, which play an important role in control statements and structures.</p><h4><strong>Control statements and structures</strong></h4><p>Control statements control what statements are executed and in which order. They can both entail conditional execution (not every piece of code gets executed) or repetition (a piece of code may get executed multiple times). For instance, the following <em>while</em> statement repeatedly executes the following two indented lines, that is, a <em>code block</em>, beneath it until the condition <em>x &gt; 0</em> becomes <em>False</em>.</p><pre>while x &gt; 0:<br>    print(x)<br>    x = x - 1  # Decrements x by 1</pre><p>If <em>x</em> is less than or equal to zero when the execution hits the <em>while</em> statement for the first time, the code block beneath it will not be executed at all.</p><p>This guide was written to help beginner programmers with their first steps in learning Python. Therefore, I have limited the guide to a minimal set of concepts including scripts, modules, statements (assignments, expressions, control statements and structures), variables, literals, operators, function calls, functions/methods, data types, and code blocks that are necessary for anyone who wants to start developing simple programs.</p><p>I have left out objects and the entire vocabulary related to them, even though Python is an object-oriented language. This is because you can and often do start programming without developing your code according to the object paradigm—I cannot think of a way to explain these concepts in a way that would not risk overwhelming a person taking his or her first steps in programming. Furthermore, because the aim is to present concepts in a way that is digestible to first-time programmers, the text glosses over many details and sacrifices some rigor to keep the guide compact.</p><p>If you spot obvious mistakes or omissions, please report those to me at aake@iki.fi</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8e078dcdb449" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to Avoid Authorship Order Disputes in Academic Publishing?]]></title>
            <link>https://aleksiaaltonen.medium.com/how-to-avoid-authorship-order-disputes-in-academic-publishing-140412e5a8c9?source=rss-394d985f3bbf------2</link>
            <guid isPermaLink="false">https://medium.com/p/140412e5a8c9</guid>
            <category><![CDATA[conflict]]></category>
            <category><![CDATA[academic-publishing]]></category>
            <category><![CDATA[authorship]]></category>
            <category><![CDATA[academic-writing]]></category>
            <category><![CDATA[dispute]]></category>
            <dc:creator><![CDATA[Aleksi Aaltonen]]></dc:creator>
            <pubDate>Fri, 05 Jul 2024 17:59:42 GMT</pubDate>
            <atom:updated>2024-07-05T17:59:42.449Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*3SLGigZX4ZFMqJ36xtlDZQ.jpeg" /><figcaption>Photo: USFWS Mountain-Prairie (CC BY 2.0)</figcaption></figure><p>Disputes over the authorship order in soon-to-be-published papers are probably the most pointless fights that can ruin personal relationships and even academic careers. Usually, there is little to win and a lot to lose by fighting over the authorship order.</p><h4>Why do fights emerge?</h4><p>The root cause for authorship disputes is that coauthoring is teamwork and there is no objective way to measure an individual contribution to a co-authored paper.</p><p>Yet, we could still try to agree on a simple, unambiguous rule on how to order author names in a paper. For instance, maybe we could simply put the names into <em>alphabetical order</em> and agree that the order is not intended to signal each authors’ relative contribution to the paper. As much as I would personally love this solution, alphabetical order is not fair. Weber (2018) reviews empirical literature on the topic and finds that there is substantial evidence that alphabetical order gives advantage to those whose name is closer to the beginning of the alphabets—i.e., if you coauthor with me and a third person, chances are that I will be ‘Aaltonen’ and you two will be ‘et al.’ whenever somebody cites the paper.</p><p>How about <em>random order</em> then? Similarly to alphabetical order, this would mean that the order of authors’ names does not carry information about the authors’ relative contribution to a paper. Yet, if there is no signal about who did the heavy lifting, this could make freeriding more attractive and reduce incentives for anyone to carry the role of a lead author. A further problem with any attempt to avoid signaling relative contributions is that single-authored papers will always signal the unique contribution of their authors.</p><p>In short, <strong>there may not be a simple ‘algorithm’ how to order authors’ names in an academic paper fairly</strong>.</p><h4>How to avoid fights, then?</h4><p>To begin with, it is good to remember that the authorship order probably matters much more to the authors themselves than others. The others probably just see your name on the paper. So, before you let the ordering of team members’ names descend into a clash of egos, ask yourself:</p><blockquote>Is it really worth it?</blockquote><p>If you feel that someone is treating you unfairly with respect to the authorship order, you are probably not going to work with them anyway in the future. Even if you win the fight, you may gain an enemy which can harm your academic prospects much more than whatever tiny advantage you may get from the ‘right’ authorship order in the paper.</p><blockquote>In short, get the paper out and move on…</blockquote><p>It may also help to try to see the situation from the perspective of the other party — especially when the dispute is between clearly more senior and more junior persons. For a junior scholar (or a PhD student), it may seem that the senior person is claiming the work that is literally the junior’s entire academic output — without putting in that much effort.</p><p>While the above may be true, it is also often true that the senior person’s input is critical in getting the paper published. Personally, I have been saved from months of wasted efforts and undoubtedly several rejections by senior scholars quickly pointing out issues in my work and suggesting how I might address them. The value of different contributions to a paper are not measured in hours worked.</p><h4>A solution</h4><p>The way I try to avoid authorship disputes is to <strong>settle with the team either: i) the authorship order, or ii) the rule by which the order will be settled, before starting to coauthor a paper.</strong> In the former option, team members can then adjust their contributions according to their ‘rank’ in the project, while the latter provides a common ground for settling the matter later on. However, it is good to record the rule in writing, for instance, in an email so that people cannot have their own recollections of the rule years later when the paper is finally coming out and the authorship order needs to be settled for good. Also, an equal contribution statement is sometimes a good way to indicate that all team members provided indispensable input to the paper throughout the process.</p><p>My solution does not guarantee that a group of scholars can agree on the authorship order, but it can help to unearth disagreements in advance, that is, before you have put a lot of effort into the project. You can then decide if you want to work on the project at all or put your precious time into something else.</p><p><em>In summary…</em></p><p><em>I have never seen anything good come out from a fight over an authorship order. The order should be settled fairly, yet there is no simple ‘algorithm’ and different team members can have different but legitimate views of what is fair with respect to a particular paper. My recommendation is to expose such differences in advance so that they can be (hopefully) resolved by reasoned argument.</em></p><p><strong>References</strong><br>Weber, M. (2018). The effects of listing authors in alphabetical order: A review of the empirical evidence. <em>Research Evaluation 27</em>(3), 238–245. <a href="https://doi.org/10.1093/reseval/rvy008">https://doi.org/10.1093/reseval/rvy008</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=140412e5a8c9" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[On the Importance (and Dangers) of Powerful Simplifications]]></title>
            <link>https://aleksiaaltonen.medium.com/on-the-importance-and-dangers-of-powerful-simplifications-0cfbc13a7955?source=rss-394d985f3bbf------2</link>
            <guid isPermaLink="false">https://medium.com/p/0cfbc13a7955</guid>
            <category><![CDATA[simplification]]></category>
            <category><![CDATA[business-school]]></category>
            <category><![CDATA[ideas]]></category>
            <category><![CDATA[knowledge]]></category>
            <category><![CDATA[parsimony]]></category>
            <dc:creator><![CDATA[Aleksi Aaltonen]]></dc:creator>
            <pubDate>Sun, 02 Jun 2024 14:57:54 GMT</pubDate>
            <atom:updated>2024-06-02T14:57:54.972Z</atom:updated>
            <content:encoded><![CDATA[<p>What is the business of business school academics? We are in the business of selling ideas on how to manage things better. While I think ours is a pretty important business, sometimes others need a little bit of convincing that we are worth the money we make from not-so-small tuition fees that most business schools charge their students.</p><p>A friend working in cancer research once asked me how does my research make the world better. For instance, how does it help to cure cancer? I replied that while his research is certainly important with respect to treating cancer patients, it alone does not cure anybody’s cancer. It is the healthcare system: people, technology (drugs, medical equipment), and organizations that cure cancer.</p><p>The knowledge that we produce in business schools can help make people, technology, and organizations deliver cancer treatments as effectively as possible to as many as possible. Simply put, <strong>our ideas matter</strong> <strong>and can change how people organize and manage organizations</strong>. Think, for instance, <a href="https://en.wikipedia.org/wiki/Frederick_Winslow_Taylor">Frederick “Scientific Management” Taylor</a> or <a href="https://en.wikipedia.org/wiki/Henry_Gantt">Henry Gantt</a> (chart) who both happen to be alumni of <a href="https://www.stevens.edu/">Stevens Institute of Technology</a> that I am joining this summer.</p><p>Yet, not every idea that comes out of a business schools is equally valuable. Some ideas are better than others or just in more demand at certain times. Currently, it is the artificial intelligence speech that everyone is selling. Some of this speech is insightful, some not—as always, the problem for the buyers of our ideas, that is, students, policymakers, and practitioners, is to tell one from the other. Ultimately, we may know only after a considerable period of time which (or whose) ideas turned out to be truly useful. However, <strong>useful ideas often come in the form of what I call a <em>powerful simplification</em>.</strong></p><p>The world is complex, even infinitely complex and to make sense of and to act upon it, we need to simplify it. Offering powerful simplifications in our publications matters because the simplifications enable others to use a study as a cognitive building block in further research or to design an intervention into a phenomenon.</p><p>Simplification does not mean that its author would suggest that the reality is as simple as a model or theory of it. Instead, a powerful simplification entails an understanding of something essential about the phenomenon and an ability to deconstruct it to its essential parts, relationships, or dynamics. This brings me to the problem I often encounter with qualitative papers I review or handle as an editor and, admittedly, have submitted to journals myself as an author.</p><p>Qualitative papers, more often than quantitative papers, may suggest that the task of research is to reveal the complexity of the studied phenomenon, or just focus disproportionately more on revealing than simplifying that complexity. Now, revealing the complexity of this or that phenomenon is not very interesting. Why? Because <strong>we can usually assume that the studied phenomenon is complex</strong>. Why else study it…</p><p>Qualitative research designs are thus more prone to dwell on the complexity of the world than quantitative designs that tend to impose a degree of simplification by design. To this end, quantitative research runs the risk of trivializing of the studied phenomenon behind fancy, scientific-looking models and techniques — but that’s a topic for another post (note that I have published both qualitative and quantitative papers in high-quality management journals so that I have some experience from the both sides of the divide).</p><p>Producing simplified conceptual knowledge is not the task of all research. For instance, a problematization of taken-for-granted knowledge can be useful for advancing knowledge in the field without offering a distinct simplification on how to intervene into a phenomenon (Alvesson and Sandberg 2011). Yet it may be useful to couch the problematization of existing knowledge in simple terms. For instance, in his paper <em>Class and Committees in a Norwegian Island Parish</em> <a href="https://en.wikipedia.org/wiki/John_Arundel_Barnes">John A. Barnes</a> (1954) struggled to make sense of a rural community as a class structure and ended up problematizing the current knowledge of the time. However, he also suggested an alternative way to label a social phenomenon as a <em>social network</em>.</p><p>Simplification, however powerful, is not the <em>only</em> value in a research paper. A narrow-minded pursuit of simplification can amount to putting the horse before the cart. As I am writing this post on a return flight from Paris to New York, let me use an analogy of an economy class seat. As a product, the economy class seat has been stripped of everything except the very necessary. It takes you safely where you need to go, but I am sure there is no person in this cabin who would not rather spend seven hours of their lives somewhere else than trapped in an economy class seat.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*KtGGAnKysnmXiuMa_olVYg.jpeg" /></figure><p>Upon landing, the crew will announce that they hope that we have enjoyed traveling with them. More accurate would of course be to hope that they have been able to minimize the torture of cheap long distance flying. This being Air France, they at least give us free champagne—probably because the company data scientists have predicted that it will make passengers feel a little bit less tortured than when flying with competitors (such as French Bee that crams one more seat per row in the same A350 plane).</p><p>The economy class seat is an apt metaphor for knowledge that often comes out from from business schools. After all, we are, at least partially, to blame for the economy class seat. Much of business school research and teaching is about how to provide value for money and how to manage business ever more efficiently, that is, how to optimize it. While this is certainly a valuable sort of knowledge and has helped to improve the standard of living massively, it is not the only kind of knowledge that successful businesses need.</p><p>A case in point is another aerospace company, Boeing. The company has arguably killed or at least seriously damaged its traditional engineering culture by relentless drive to cut costs and maximize profits—no doubt with the help of an army of professional managers and their learnings from business school academics (see John Oliver from 8:13 in <a href="https://youtu.be/Q8oCilY4szc?si=brMVnGN2IwNV_e5I">https://youtu.be/Q8oCilY4szc?si=brMVnGN2IwNV_e5I</a>). Just like a company that can only see ‘shareholder value’ or ‘cost efficiency’, if business school academics can only produce powerful simplifications without regularly stepping back and reflecting why and for what purposes we simplify the world, our importance is bound to diminish in increasingly datafied business environments where engineers can probably soon program such tasks into machines.</p><p>Instead of trying to be (or to become) narrow ‘management engineering’, management scholarship should keep cherishing its diverse roots in social sciences to remain relevant in the long run. We need those metaphorical economy class seats and even business class seats but also reflections on air travel itself, entire transportation systems, why and for what purpose we travel.</p><p><strong><em>Postscript. </em></strong><em>My argument in this post is, of course, nothing but a simplification in itself. In some ways, I have been merely reiterating the old idea of </em><a href="https://en.wikipedia.org/wiki/Occam%27s_razor"><em>Occam’s razor or parsimony</em></a><em> by giving it a spin in a particular context.</em></p><p><strong>References<br></strong>Alvesson, M., &amp; Sandberg, J. (2011). Generating research questions through problematization. <em>Academy of Management Review, 36</em>(2), 247–271. <a href="https://doi.org/10.5465/AMR.2011.59330882">https://doi.org/10.5465/AMR.2011.59330882</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=0cfbc13a7955" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[I Do Better Research than You — On Respect and Disagreement in Academia]]></title>
            <link>https://aleksiaaltonen.medium.com/i-do-better-research-than-you-on-respect-and-disagreement-in-academia-dadfb57b7302?source=rss-394d985f3bbf------2</link>
            <guid isPermaLink="false">https://medium.com/p/dadfb57b7302</guid>
            <category><![CDATA[disagreement]]></category>
            <category><![CDATA[respect]]></category>
            <category><![CDATA[academia]]></category>
            <category><![CDATA[methodology]]></category>
            <dc:creator><![CDATA[Aleksi Aaltonen]]></dc:creator>
            <pubDate>Fri, 19 Apr 2024 13:18:09 GMT</pubDate>
            <atom:updated>2024-04-20T19:17:03.403Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*wEitFMYGp2wyZ7ZvYOpcJw.png" /></figure><h3>I Do Better Research than You — On Respect and Disagreement in Academia</h3><p>Academic researchers are often proud of their research and, in particular, <em>how</em> they do their research: my combination of theory, method, and epistemology is a little bit (if not a lot) better than yours.</p><blockquote>Of course, I would have not chosen my approach unless I think it is the Right way of studying things!</blockquote><p>The thing is that <strong>neither</strong> <strong>you nor me made an informed choice how we do our research.</strong> Instead, the way we do our research was likely handed down to us by the postgraduate program where we ended up, for whatever reason, doing our doctorates. To have originally made an informed decision about a personal approach to research is virtually impossible. It would require one to possess a solid understanding of different options before making a choice — which is only possible if you first learn and apply the different ways of doing research…</p><p>This is not to say that certain types of research would not be more popular, useful, or even better according to some criteria in different fields and at different times. For instance, I was trained to do qualitative case studies. Yet, as I saw the rising opportunities in the vast amounts of data available in the digital environment, I trained myself in computational and econometric methods to be able to tap those opportunities. As a result, I have published positivistic quantitative papers (Aaltonen and Seiler 2016, Shi et al. 2023), qualitative papers drawing on critical realism (Aaltonen and Tempini 2014) and on performativity (Aaltonen and Stelmaszak 2023), and papers that do not neatly fall into any such category (Kallinikos et al. 2013). In other words, I have done research in ways that are seemingly incompatible or even mutually exclusive.</p><p>You may call me a methodological agnostic (some may think <em>heretic</em>). I am less interested in committing to a certain methodology than understanding phenomena. The benefit of working with totally different methods is, however, that it allows to learn something about methodology in general. Most importantly, I have learned that:</p><blockquote>Just because I don’t understand research done using a particular approach does not make it inferior to my research.</blockquote><p>This may sound obvious — I would have probably agreed with such a noble statement even before I learned to do different types of reseach. However, the true test is if I can respect a different type of research in practice, especially when it comes to reviewing a paper, hiring a new colleague, or assessing a promotion case. Respect does not mean that I should pretend that I find different types of research equally valuable; indeed, I think lots of people are wasting their lives doing incredibly boring research. Yet, I must accept that they have the right to think the same about my research because:</p><blockquote>Ultimately, we don’t know whose research is going to be truly impactful over the years, yours or mine — all we can do is to give our best shot while letting others do the same.</blockquote><p>So, going back to my initial provocation, do I really think that my approach is the Right approach? Yes, in the sense that I think it works very well for me given my experience, interests, and skills that I have developed over the years. No, in the sense that I don’t think everyone should do research like me. The beauty of academic work is that we don’t have to agree on the Right way of doing research as long as we respect each other.</p><p>In fact, I have benefited tremendously from working with colleagues who think differently from me. Nothing exposes problems in my own thinking better than having to convince a bunch of sceptic colleagues who want to disagree with me. I recently had an interesting discussion with Nick Berente about the importance of disagreement or what Nick calls ‘adversarials’ in academic knowledge production. Knowledge does not grow when we agree with each other. Instead,</p><blockquote>Knowledge grows when we disagree with each other and, in this sense, functioning academia is an institutionalized disagreement.</blockquote><p>Unfortunately, disagreement among academics is not always respectful. It can be hostile. Hostility can also force us to improve our research and to strengthen our arguments, but I am not sure if it is a <em>more</em> effective way to improve our thinking than respectful disagreement. I hope not and I believe that the foundation for mutual respect can be found in the recognition of the following statement:</p><blockquote>However strongly you may believe in the way you do your research, you cannot deny that other ways of doing research have also produced insightful findings and useful knowledge.</blockquote><p><strong>Cited papers</strong></p><p>Aaltonen, A., &amp; Seiler, S. (2016). Cumulative growth in user-generated content production: Evidence from Wikipedia. <em>Management Science</em>, 62(7), 2054–2069. <a href="https://doi.org/10.1287/mnsc.2015.2253">https://doi.org/10.1287/mnsc.2015.2253</a></p><p>Aaltonen, A., &amp; Stelmaszak, M. (2023). The performative production of trace data in knowledge work. <em>Information Systems Research</em>, Articles in Advance. <a href="https://doi.org/10.1287/isre.2019.0357">https://doi.org/10.1287/isre.2019.0357</a></p><p>Aaltonen, A., &amp; Tempini, N. (2014). Everything counts in large amounts: A critical realist case study on data-based production. <em>Journal of Information Technology</em>, 29(1), 97–110. <a href="https://doi.org/10.1057/jit.2013.29">https://doi.org/10.1057/jit.2013.29</a></p><p>Kallinikos, J., Aaltonen, A., &amp; Marton, A. (2013). The ambivalent ontology of digital artifacts. <em>MIS Quarterly</em>, 37(2), 357–370. <a href="https://doi.org/10.25300/MISQ/2013/37.2.02">https://doi.org/10.25300/MISQ/2013/37.2.02</a></p><p>Shi, R., Aaltonen, A., Henfridsson, O., &amp; Gopal, R. D. (2023). Comparing platform owners’ early and late entry into complementary markets. <em>MIS Quarterly</em>, 47(4), 1727–1744. <a href="https://doi.org/10.25300/MISQ/2023/17413">https://doi.org/10.25300/MISQ/2023/17413</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=dadfb57b7302" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What Makes a Good Hypothesis?]]></title>
            <link>https://aleksiaaltonen.medium.com/what-makes-a-good-hypothesis-e22d08b7b90e?source=rss-394d985f3bbf------2</link>
            <guid isPermaLink="false">https://medium.com/p/e22d08b7b90e</guid>
            <category><![CDATA[hypothesis]]></category>
            <category><![CDATA[theory]]></category>
            <category><![CDATA[data]]></category>
            <category><![CDATA[research-paper]]></category>
            <dc:creator><![CDATA[Aleksi Aaltonen]]></dc:creator>
            <pubDate>Wed, 13 Mar 2024 13:52:22 GMT</pubDate>
            <atom:updated>2024-03-13T13:52:22.154Z</atom:updated>
            <content:encoded><![CDATA[<p>Not every research paper need hypotheses but when you have them, they better be good. <em>“Hypotheses are dead simple so writing them cannot be too difficult”</em>, I thought for many years before starting to work on research designs that rely on hypotheses and realizing how wrong I had been.</p><p>In this post, I summarize my lessons learned about the important role of hypotheses in a research paper. To this end, I acknowledge that this is not a post about (null) hypothesis testing in statistics; also, my views are undoubtedly shaped by my publishing in a particular academic field, that is, Management Information System (but so are yours, too).</p><h4><strong>The Purpose of Hypotheses</strong></h4><p>To begin with, the purpose of hypotheses in a research paper is to define specific tests that can be unambiguously answered yes or no based on empirical analysis. Well, nothing is ever completely unambiguous so let’s say relatively unambiguously. For instance, let us look at a hypothesis from a paper that I am currently working on:</p><blockquote>H1: Code correctness increases programming solution quality.</blockquote><p>The hypothesis is made of two constructs, ‘code correctness’ and ‘programming solution quality’, and a proposed positive relationship from the former to the latter. By measuring the constructs and using, for instance, some statistical technique we can test if the proposed relationship holds or not. However, having been trained as a qualitative researcher, it took me many years to understand what’s the point in testing such an utterly simplistic statement — because <em>it always depends</em>.</p><p>Of course it depends! But, the point is to design the empirical study around the hypotheses so that everything that it depends on have been accounted for as much as possible. In other words, in a hypothesis-driven study, a successful research design enacts a laboratory, whether imaginary or real, in which it is possible to say something interesting by testing simple hypotheses.</p><p>The H1 above may also sound too obvious to be worth even tested. This is probably because the hypothesis <em>is </em>obvious apart from how exactly one should measure the two constructs. Correct code (in this case, defined as program code that executes without errors) should obviously increase the quality of a programming solution. Yet, in the paper, the hypothesis serves as a baseline check that the overall research model and our data aligns with what we know about the studied phenomenon in general. As such, the point is not so much to find supporting evidence for H1 but to <em>fail</em>to falsify it in the sprit Karl Popper.</p><p>Yet, not every simplistic statement makes a good hypothesis…</p><h4>Hypotheses Are Not Empirical Speculations</h4><p>Based on my experience as a reviewer and journal editor, a common mistake is to write hypotheses as empirical speculations, that is, as more or less educated guesses what the data might show. This may perhaps make a paper look more ‘scientific’ to an uninformed reader, but such hypotheses are usually nothing but unnecessary decorations. If your analysis does nothing more than describes what’s there in the data, you can do that without convoluting the description with hypotheses.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6TnNUCH8c85vtLArHN9qAw.png" /><figcaption>Figure 1. A hourglass model of hypotheses.</figcaption></figure><p>By contrast, Figure 1 illustrates the idea that <strong>useful hypotheses function as the pivotal element between data analysis and theory</strong> in a research paper. By testing the hypotheses, we should learn something theoretically interesting about the studied phenomenon. This means that hypotheses must both derive from theory and be functional elements in the empirical analysis, which is easier said than done. They entail pinning down a theoretical dilemma in an empirically operational manner, which makes the hypotheses the crux of the analysis despite often appearing as simple one-line statements.</p><h4><strong>Hypotheses Are Short Written Statements that Connect Theory with Evidence</strong></h4><p>Good hypotheses are short statements that are crafted with extreme care. One should not think them as abstract ideas but take them literally as written statements. Every word matters and should be chosen to serve the critical link between theoretical ideas and the analysis of empirical data in the best possible way. I find myself typically writing and rewriting every hypothesis several times to find as good wording as possible.</p><p>Common mistakes in writing hypotheses include merging several empirical tests into one hypothesis. If the hypothesis cannot be settled with a specific empirical test, it should be broken down into several hypotheses. Also, one should not include boundary conditions to the hypotheses for this quickly makes the hypotheses unwieldy and you can never include all the conditions.</p><h4>Concluding Remarks</h4><p>I hope this post offers some useful reflections from my journey to understanding the role of hypotheses in a research paper. These can be summarized into the following three statements:</p><ol><li>A good hypothesis defines a single empirical test.</li><li>A good hypothesis links empirical data to theory.</li><li>A good hypothesis is relatively short.</li></ol><p>Finally, I stress that the reflections in this post are based my own gradual learnings from working with hypotheses in research papers. As such, they are not particularly rigorously argued and there is a lot more to hypotheses than what has been said here. For instance, there is an important question whether study hypotheses must be fixed before collecting and analyzing data, or can the hypotheses be adjusted along with the analysis. People have different views about this that are rooted in the nature of knowledge in different academic disciplines. As I asked a friend who is trained in statistics to comment a draft version of this post, she told me that <em>“a statistician would kill you over even suggesting that the hypothesis can be adjusted along the analysis”.</em> Clearly, this would a worthy topic for another post!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e22d08b7b90e" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Data-driven, Data-based or Data Innovation?]]></title>
            <link>https://aleksiaaltonen.medium.com/data-driven-data-based-or-data-innovation-c7e3f0c22cae?source=rss-394d985f3bbf------2</link>
            <guid isPermaLink="false">https://medium.com/p/c7e3f0c22cae</guid>
            <category><![CDATA[innovation]]></category>
            <category><![CDATA[data-driven]]></category>
            <category><![CDATA[data]]></category>
            <dc:creator><![CDATA[Aleksi Aaltonen]]></dc:creator>
            <pubDate>Mon, 12 Feb 2024 14:02:45 GMT</pubDate>
            <atom:updated>2024-02-12T14:02:45.742Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*CXaAIGfZ1M0LKh_vS1uc1Q.jpeg" /><figcaption>Photo by <a href="https://www.flickr.com/photos/jdhancock/4617759902">JD Hancock</a> (<a href="https://creativecommons.org/licenses/by/2.0/">CC BY 2.0 DEED</a>)</figcaption></figure><p>Data are the building blocks of the ongoing digital revolution. They fuel incremental and breakthrough innovations to organizational process, products and services, and digital data underpin whole new ways of organizing productive activities.</p><p>Yet, we seldom stop and think what data actually are or exactly how data perform their critical role as the key resource and medium of organizing in contemporary economy. This is reflected in how we talk about data-driven innovation using ambiguous terminology that lacks resolution even at the most general level. The following is a simple typology for thinking clearly about innovations related to digital data.</p><p>First, innovation maybe <strong>data-driven</strong> in the sense that data analytics and artificial intelligence are used, for instance, to segment the market in a new way to increase the effectiveness of product development. For example, Netflix may study its subscribers viewing patterns to identify opportunities to serve as large portion of its viewers as possible with a new type of series.</p><p>Second, innovation may also be <strong>data-based</strong> in the sense that a new product or service is essentially fashioned out of data (Aaltonen et al., 2021). For instance, advertising audiences, institutional rankings, credit scores, and popularity metrics are literally measured into existence and would not exist without data. In contrast to data-driven innovation (such as the TV series in the example above), data-based innovation would not be possible without data.</p><p>Third, it is important to recognize that data itself can be an innovation. A <strong>data innovation</strong> is <em>“a new valuable way to render phenomena as data that can be processed computationally”</em> (Aaltonen and Penttinen, 2021, pp. 5928–5929). For instance, it has taken twenty years to develop eXtensible Business Reporting Language (XBRL) that allows the reporting of company financial statements in a machine-readable format. New types of valuable data do not appear by accident but need to be often painstakingly created.</p><p>The above typology is admittedly crude and the distinctions between the categories may not always be clear cut — a particular innovation can incorporate features from all the three types of data-related innovations. Even so, the typology shows that is possible and, indeed, I would argue critical to make more nuanced distinctions to understand the role of digital data in contemporary innovation.</p><p>REFERENCES</p><p>Aaltonen, A., Alaimo, C., &amp; Kallinikos, J. (2021). The making of data commodities: Data analytics as an embedded process. <em>Journal of Management Information Systems,</em> <em>38</em>(2), 401–429. <a href="https://doi.org/10.1080/07421222.2021.1912928">https://doi.org/10.1080/07421222.2021.1912928</a></p><p>Aaltonen, A., &amp; Penttinen, E. (2021). What makes data possible? A sociotechnical view on structured data innovations. <em>Proceedings of the 54th Hawaii International Conference on System Sciences</em>, USA, 5922–5931. <a href="http://hdl.handle.net/10125/71336">http://hdl.handle.net/10125/71336</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c7e3f0c22cae" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[All Data Are Wrong, but Some Are Useful]]></title>
            <link>https://aleksiaaltonen.medium.com/all-data-are-wrong-but-some-are-useful-5920b22eb89f?source=rss-394d985f3bbf------2</link>
            <guid isPermaLink="false">https://medium.com/p/5920b22eb89f</guid>
            <category><![CDATA[data]]></category>
            <category><![CDATA[gender]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[complexity]]></category>
            <category><![CDATA[representation]]></category>
            <dc:creator><![CDATA[Aleksi Aaltonen]]></dc:creator>
            <pubDate>Thu, 18 Jan 2024 14:21:54 GMT</pubDate>
            <atom:updated>2024-04-03T15:59:48.379Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/1*WTFsT2sS1O-LuWHzQU18ug.jpeg" /></figure><p>Algorithms and complex models have been all the rage across management disciplines for some time. Yet, data matters just as much: <em>if you put a turkey in, it’s not going to come out as an eagle — </em>however much you cook the data with your algorithms.</p><p>The problem is that you can never get data completely right. The moment you capture a phenomenon as data, you impose artificial distinctions on reality. Some distinctions are better than others and capture perhaps more imporant aspects of the phenomenon of interest but, paraphrasing the old saying:</p><blockquote>All data are wrong, but some are useful.</blockquote><p>Data are often treated as an unproblematic, ready-to-use resource measured by its volume. Such an assumption may be necessary and justified in the context of a particular research effort, but it should not obscure the fact that data are a much more complex matter than implicit assumptions embedded in our statistical methods suggest.</p><p>Against this background, it is not surprising that critical scholarship on data typically fault one or another information system for not representing the world ‘as it really is’. This position rightly calls for attention to important nuances and omissions with any data or representation, but fails to acknowledge a hard limitation of computational technology:</p><blockquote>It is impossible to represent the world as it really is, in its full richness, as data.</blockquote><p>If we will only accept systems (or statistical analyses) that represent, for instance, certain people perfectly, we will never build any system. It is always possible to try make the collected data more detailed and accurate, but the price of more nuanced and in this sense better data is increasing complexity. Complexity has a very real and often escalating cost to any effort to build an application in practice. Let me give you an example.</p><p>It is easier to work with a binary male vs. female variable than a multilevel classification in which diverse ways to define gender identity are not mutually exclusive. The latter will undoubtedly allow capturing the gender identity of the diversity of people better, but it also imposes practical burden on the production and use of the data. We could even think of allowing people to describe their gender identity freely as text. This would provide maximum flexibility in capturing people’s gender identities, but also largely defeat the purpose why we typically collect data. That is, to classify things so that we can work with a limited number of categories. If everyone are allowed to create their own category, there are no categories at all.</p><blockquote>Even if we can never have perfect data, we can strive for better data.</blockquote><p>However, not being able to produce perfect data should not be used as an excuse for not striving for better data. We have increasingly powerful computational systems that enable us to contain much more complexity than before and, consequently, allow to add detail to data where it matters the most. That is, the price of complexity we have to pay for better data does not stay the same but can decrease with the rapidly evolving technology. We must use this opportunity to continuously improve how we capture important aspects of reality in data, and to minimize the violence that our data does to those who might not fit into traditional ways of categorizing the world.</p><p><em>This post is based on the notes I wrote for my opening statement in the “Algorithmic Bias and Data Injustice: Dark Side or Dark Matter?” panel symposium in the 83rd Annual Meeting of the Academy of Management (AoM) on August 7, 2023, Boston, MA.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5920b22eb89f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[NYT vs OpenAI and Microsoft: Three Scenarios with a Big Impact on Our Digital Environment]]></title>
            <link>https://aleksiaaltonen.medium.com/nyt-vs-openai-and-microsoft-three-scenarios-with-a-big-impact-on-our-digital-environment-7fff4c233be7?source=rss-394d985f3bbf------2</link>
            <guid isPermaLink="false">https://medium.com/p/7fff4c233be7</guid>
            <category><![CDATA[training-data]]></category>
            <category><![CDATA[openai]]></category>
            <category><![CDATA[lawsuit]]></category>
            <category><![CDATA[copyright]]></category>
            <category><![CDATA[fair-use]]></category>
            <dc:creator><![CDATA[Aleksi Aaltonen]]></dc:creator>
            <pubDate>Wed, 03 Jan 2024 13:50:48 GMT</pubDate>
            <atom:updated>2024-01-03T13:50:48.305Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*yu-1BTvSp3y9gdERIg-bYg.jpeg" /></figure><p><strong>It was bound to happen.</strong> On December 27, 2023, The New York Times sued OpenAI and Microsoft for mass copyright infringement related to how the companies used the publisher’s content as training data for their generative artificial intelligence systems. The lawsuit is the latest and the most high-profile case in a string of similar lawsuits that have been filed against AI companies recently.</p><p>According to the <a href="https://nytco-assets.nytimes.com/2023/12/NYT_Complaint_Dec2023.pdf">case documents</a>, the The New York Times wants OpenAI and Microsoft to stop using its content as training data and to destroy the content from their systems along with the models trained on the content. This would not only require re-training systems such as ChatGPT with potentially poorer training data but also open the door for other copyright owners to require OpenAI and Microsoft to do the same.</p><p>The case is built on two arguments.</p><p>First, The New York Times argues that the defendants have already benefited commercially and aim to further benefit from derivative products unlawfully based on the copyrighted content. In other words, the financial interests of the company and its shareholders are being violated. Second, The New York Times argues that its capacity to fund high quality journalism is at stake and, more generally, journalism in general if technology companies are allowed to scrape and use copyrighted content under the fair use doctrine.</p><p>A lot is at stake in the lawsuit that may turn out to be a watershed moment in the evolution of AI. Generative AI is just as much about the data than the algorithms.</p><p>In some ways, the situation is analogous to the rise of search engines twenty years ago. Search engines indexed everyone’s content and started selling advertisements next to the index, rerouting massive amounts of advertising money away from publishers. Yet, search engines by design pass visitors on to publishers’ websites. Generative AI tools are not built to do this, which is why they can have potentially much more dramatic impact on publishers. However, this time publishers such as The New York Times are better prepared to protect their interests.</p><p>Below, I sketch three scenarios how the lawsuit may unfold.</p><h4><strong>OpenAI and Microsoft win</strong></h4><p>This seems an unlikely outcome. People who have read through the case documents (including me) against OpenAI and Microsoft seem to think that the case is strong. I would also imagine that OpenAI and Microsoft will prefer, if possible, an out-of-court settlement to the risk of losing the case. Yet, when it comes to complex court cases, anything is possible. If OpenAI and Microsoft win the case, this would probably redefine the idea ‘fair use’ of copyrighted content quite a bit. Systematically harvesting millions of copyrighted documents and using them to build a competing product cannot be fair use, or can it?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*KxgzKwAbxQuuzSXJoYypkQ.png" /><figcaption>Platforms such as Squarespace already allow to stop AI from scraping your content.</figcaption></figure><p>Yet, there is a caveat. Even if OpenAI and Microsoft would win, no one can force copyright owners to make their content available for technology companies — and technology can be used to stop AI from scraping content. Some content platforms have already implemented such features and the victory for the two tech companies could trigger an arms race to secure content ecosystems against unwanted scraping by the developers of AI.</p><h4><strong>Out-of-court settlement</strong></h4><p>This is the most likely outcome if The New York Times is more interested in its shareholders than journalism as a public good. Indeed, given that The New York Times was in talks with OpenAI and Microsoft before suing them, a possible interpretation of the lawsuit is that it is a negotiation tactic. The New York Times will use it as a leverage to reach a favorable agreement with the companies. This way, the company could protect its <em>own</em> capacity to produce high quality journalism.</p><p>Other large corporations will then undoubtedly follow, using armies of lawyers to erect a complex web of contracts with AI companies to supply them training data — while smaller players and individuals are left out in the cold. Axel Springer has already entered into such a relationship with OpenAI. All in all, I would not be surprised if big copyright holders will find a way to retain status quo by forcing OpenAI and Microsoft to share some of the revenues from products based on generative AI. Whether anything will trickle down to individual content producers will remain to be seen (I would not hold my breath).</p><h4><strong>The New York Times wins</strong></h4><p>If The New York Times is interested in journalism as a public good, then it might see through the lawsuit and refuse an out-of-court settlement. If the publisher wins, this will not be the end of generative AI or Microsoft or even OpenAI. The companies will have to cough up some damages and maybe retrain their models and, most importantly, come up with solutions to source training data fairly. If technology companies can create highly sophisticated generative AI systems, they should be able come up with solutions to distribute value created by AI tools to those whose works are used as training data. A win for The New York Times could set the development of generative AI on a trajectory where technology companies have an incentive to do this.</p><p><em>Finally, OpenAI and Microsoft probably expected to be sued for their use of copyrighted material as training data so it will be interesting to see what kind of approach they take to defend themselves. Is it going to be based on ‘fair use’ argument as the case documents assume or something else? One way or the other, The New York Times vs. OpenAI and Microsoft case will have a major impact on the development of generative AI and our future digital environment.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7fff4c233be7" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Algorithms, Humans, and the Duality of Data–A Futile Attempt to Define IS Phenomena]]></title>
            <link>https://aleksiaaltonen.medium.com/algorithms-humans-and-the-duality-of-data-a-futile-attempt-to-define-is-phenomena-a12c8d28573c?source=rss-394d985f3bbf------2</link>
            <guid isPermaLink="false">https://medium.com/p/a12c8d28573c</guid>
            <category><![CDATA[data]]></category>
            <category><![CDATA[discipline]]></category>
            <category><![CDATA[information-systems]]></category>
            <category><![CDATA[humans]]></category>
            <category><![CDATA[algorithms]]></category>
            <dc:creator><![CDATA[Aleksi Aaltonen]]></dc:creator>
            <pubDate>Thu, 31 Aug 2023 02:27:01 GMT</pubDate>
            <atom:updated>2024-04-03T16:09:00.499Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*b-5vfQ2xv17HsY7qa-5TNw.png" /></figure><p>My PhD supervisor Jannis Kallinikos once said that a <em>field</em> of research is defined by the phenomenon it studies, while an academic <em>discipline</em> constructs its object of research. By this definition, the study of information systems is merely a field, even if we as IS scholars often like to call ourselves a discipline...</p><p>I have a huge respect for academic disciplines such as economics but, at the same time, I am happy to work on a field that is in a perpetual and, indeed, seemingly hopeless chase of its phenomena. If you ask from ten IS scholars what makes something a study of information systems, you probably get ten different answers. Here is mine:</p><blockquote>Information systems studies phenomena that are always, in one way or another, about algorithms, humans, and data.</blockquote><p>Let me elaborate.</p><p>All digital information systems are made of <strong>algorithms</strong>. The study and development of algorithms as technological entities belong primarily to computer scientists and engineers. There are great scholars (and friends!) in the field of information systems who develop algorithms, but as a field our advantage is typically to put algorithms in context. For instance, the study of algorithmic management strives to understand how management and work changes when managerial action is taken by algorithmic procedures.</p><p>While algorithms are certainly a very important topic to study, there is a downside to the rise of evermore powerful algorithms and models. They tend to summon vague and often naive talk about ‘autonomous’ systems as if digital systems would be at the verge of taking a life on their own. Yet, no computational system is autonomous in the sense of operating independent of <strong>humans</strong>. Take, for instance, generative artificial intelligence. The generation of text and images is triggered by human prompting, not to mention that the systems are designed, trained, and tuned by humans. The freely available version ChatGPT has been trained with data until September 2021 — unless some humans decide to feed the system with more training data, the system will soon start showing signs of serious dementia.</p><p>Finally, <strong>data</strong> are the world in which algorithms live. Data constitute the fundamental link between the social world of humans and algorithmic procedures. In this sense, data are both technological entities, digital material, so to speak, and semantic entities that can be used to make sense of something. This duality makes data a particularly interesting element of information systems. Any attempt to use digital data are constrained by the way the data are designed, produced, and stored in practice, yet data also need to be embedded in specific human practices and institutions to function as data.</p><h4>So what?</h4><p>The trinity of algorithms, humans, and data usefully highlights the importance of keeping in mind that all three spheres are always involved. Actual research efforts often emphasize one sphere over the others, which is fine as long as we do not deny the existence of the others.</p><p>For instance, better awareness of all three spheres could help putting zombie concepts like ‘autonomous systems’ out their misery (1), or act as a reminder to those who wish take into consideration the interpretive flexibility of technology (or any of its numerous descendants) that any interpretation needs something to interpret, and that that something constrains the interpretation (otherwise one is not interpreting anything but merely fantasizing).</p><p>On a more personal level, to me it is interesting how data sits in-between the sphere of humans and the sphere of algorithms, highlighting what could be called <em>the duality of data</em>: digital data are always technological and social matter. To this end, the inability of information systems research to develop a strong disciplinary core can be seen as its strength — the field has remained open to bringing together the understanding of storage technologies, conceptual modeling and semiotics, and social practices in a way that allows us to do cutting edge research into some of the most exciting contemporary data-based phenomena.</p><p><em>Footnotes</em></p><p>(1) A zombie concept is an idea that refuses to die despite being repeatedly refuted. See, Peters, B. G. and Nagel, M. L. 2020. <a href="https://www.cambridge.org/core/elements/abs/zombie-ideas/77F6FA03E416DB0E578213464BA1E860">Zombie Ideas: Why Failed Policy Ideas Persist</a>. Cambridge University Press.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a12c8d28573c" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>