<?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 Chris Michael on Medium]]></title>
        <description><![CDATA[Stories by Chris Michael on Medium]]></description>
        <link>https://medium.com/@cjemichael?source=rss-45e88a079b85------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*rGCv66UPcpPH8zE0.jpeg</url>
            <title>Stories by Chris Michael on Medium</title>
            <link>https://medium.com/@cjemichael?source=rss-45e88a079b85------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 02:50:24 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@cjemichael/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[Open Banking in South Africa]]></title>
            <link>https://medium.com/@cjemichael/open-banking-in-south-africa-798572b0ae0d?source=rss-45e88a079b85------2</link>
            <guid isPermaLink="false">https://medium.com/p/798572b0ae0d</guid>
            <category><![CDATA[fintech]]></category>
            <category><![CDATA[psd2]]></category>
            <category><![CDATA[open-banking]]></category>
            <dc:creator><![CDATA[Chris Michael]]></dc:creator>
            <pubDate>Wed, 10 Oct 2018 14:04:43 GMT</pubDate>
            <atom:updated>2018-10-10T14:04:43.859Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6GWks9BULuw1RqpXjOqq3w.jpeg" /></figure><p>Last week I meet several banks in South Africa, courtesy of <a href="https://www.pingidentity.com/en.html">Ping Identity</a> and <a href="https://www.ubusha.co.za/">Ubusha</a>. Over the course of two days, I spoke to <a href="https://www.nedbank.co.za/">Nedbank</a>, <a href="https://www.absa.co.za/">Absa</a> and <a href="https://www.standardbank.com/">Standard Bank</a>, as well as a number of challengers.</p><p>I was interested to see what they thought of the work the <a href="https://www.openbanking.org.uk/">Open Banking Implementation Entity</a> (OBIE) has been doing in the UK to develop a standard and certification process, and how this could help these banks meet regulatory and/or market needs regarding Open Banking APIs.</p><h3>Open Banking will deliver significant customer benefit</h3><p>Open Banking APIs should offer a more secure, reliable alternative to screen scraping. This will reduce costs and increase the opportunities for money management tools, helping personal and business customers make more of their money.</p><p>I believe the most exciting potential is in payments — to offer a better, lower cost, more reliable, faster, more transparent alternative to card schemes. For this potential to be realised, banks need to implement APIs which meet both regulatory and market needs.</p><h3>Introducing the OBIE standard</h3><p>Version 3 is a truly global standard, which should allow any bank in Europe and beyond to build a great API interface which complies with regulations (e.g. the CMA Order and PSD2).</p><p>The <a href="https://www.openbanking.org.uk/providers/standards/">OBIE standard</a> is the result of a massive collaboration over two years, involving thousands of stakeholders, including banks, third parties, vendors, consultancies and other standards bodies. All have been able to contribute in an open forum online and in person — over 50 workshops to date and counting.</p><p>The security model has been developed in conjunction with the <a href="https://openid.net/">Open ID Foundation</a>, and the standard has now adopted the FAPI (Financial Grade API) and CIBA (Client Initiated Backchannel Authentication) profiles. This allows for great customer experience and best in class security to protect banks, third parties and end customers. OBIE has also committed to supporting Dynamic Client Registration, which enables third parties to onboard automatically with banks, reducing the friction, risk and cost of manual processes.</p><p>The standard also includes comprehensive <a href="https://www.openbanking.org.uk/wp-content/uploads/Customer-Experience-Guidelines.pdf">Customer Experience Guidelines</a> which define what banks and third parties must do (to meet regulatory requirements) and should do (to meet market/best practice needs) when authenticating a customer’s access to their bank account.</p><p>The largest banks in the UK have been live since January and third parties have been connecting and going live with real customers. As of September 2018, these banks are now live with v2 of the standard, and many more are actively building their PSD2 interface using v3.</p><p>The OBIE has committed to a roadmap well into 2019, which will add even more functionality and value to the standard.</p><p>So, in summary, this is already a proven standard, which is getting better all the time.</p><h3>Going beyond the standard</h3><p>However, there have been a number of documented issues, and it still too early to see end-user benefits at scale. OBIE are doing several things to address this:</p><p>1. They have developed a suite of <a href="https://openbanking.atlassian.net/wiki/spaces/DZ/pages/23856067/OB+OIDC+Conformance+Suite">conformance tools and a certification process</a>. This is currently focused on the security profile, but OBIE are extending this to cover the functional APIs and Customer Experience Guidelines. This will be a massive benefit to banks (to help them test and prove compliance to the standard), regulators (to help them evaluate whether banks are compliant), and also third parties (so they can see which banks have implemented correctly).</p><p>2. OBIE have also developed a trust framework for banks and third parties. This is called the <a href="https://www.openbanking.org.uk/providers/directory/">Directory</a>, and includes a number of features which reduce barriers and risks for banks and third parties. The Directory has been live since late 2017, and the next version will meet all PSD2 requirements, including compatibility with eIDAS certificates.</p><p>3. Both OBIE and third party developers provide reference code libraries, including a model (sandbox) banks. This has been essential to speed up development and adoption.</p><p>4. Finally, OBIE provide additional support services to help banks and third parties test and connect with each other, before, during and after go live.</p><p>As an implementation entity (rather than just a standards body), there is much more OBIE are doing in this space. Ultimately the goal is to drive adoption of the Open Banking ecosystem.</p><h3>Open Banking in South Africa</h3><p>Unlike the UK and Europe, the regulators in South Africa have not yet mandated Open Banking. Personally I do not believe Open Banking will work without some regulation. Customers need assurance when they use a third party provider, that they are dealing with a regulated entity. They also need a means to recourse in the event something goes wrong.</p><p>Does this mean that banks in South Africa should be ‘forced’ by the regulator to provide unlimited API functionality and access for free? Or is it enough to just create some basic rules and leave the rest to market forces?</p><p>I don’t know the answer, but maybe it’s somewhere in the middle. In my experience, companies produce better products when these are aligned to their core business objectives and when they have a financial incentive. I am also convinced that having a great API should be at the heart of any bank’s strategy, so that they can use this to build partnerships with third parties who in turn can bring in more customers and more revenue.</p><p>Perhaps the biggest incentive for any bank in any market is the move that big tech giants are making in this space. They will bring customers at scale and will want APIs which offer a great user experience, great functionality, and are fast, reliable and highly available. They also want these APIs to be secure and built to a well-documented standard. And while they may not be knocking on the door of South African banks just yet, it is only a matter of time.</p><h3>Next steps?</h3><p>I would urge all banks in any market where there is not yet a regulatory mandate for Open Banking to get on the front foot and build a great API. Banks who do this will realise several benefits:</p><p>1. A competitive advantage, in particular by developing (commercial) relationships with third parties ahead of other banks.</p><p>2. Reduced likelihood of being forced by a regulator to do this for free and/or in short timeframes.</p><p>3. Reduced delivery risk, in the event the regulator does step in, since much of the required functionally may already be delivered.</p><p>Banks who follow a standard will find it easier to build their API, as this will reduce the cost and risk of reinventing the wheel. Since OBIE have done the hard work here, I would strongly urge all banks to start by looking at this standard.</p><p>Banks should pay particular attention to the developer community, whether small start-ups or larger, more established third parties. Making the API is well documented and easy to use is just the start. Banks should also provide sandbox implementations and get feedback via events and hackathons.</p><p>Finally, Identity is key to delivering Open Banking APIs securely and with a great user experience. To do this, banks need the right technology, especially in the Identity and Access Management space. Ping Identity has a proven track record with many banks in the UK, Europe and Globally. And Ubusha have the expertise and local market presence to help banks in South Africa implement and integrate IAM solutions and hence deliver Open Banking APIs.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=798572b0ae0d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Will Artificial Intelligence end humanity?]]></title>
            <link>https://medium.com/@cjemichael/will-artificial-intelligence-end-humanity-350a0365eddf?source=rss-45e88a079b85------2</link>
            <guid isPermaLink="false">https://medium.com/p/350a0365eddf</guid>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[deep-learning]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Chris Michael]]></dc:creator>
            <pubDate>Fri, 07 Oct 2016 09:19:46 GMT</pubDate>
            <atom:updated>2016-10-07T09:48:12.697Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*WBNwAlkHjSwpnVqH9z3yRw.jpeg" /><figcaption>Image credit: <a href="http://thepookieprotest.wikia.com/">http://thepookieprotest.wikia.com</a></figcaption></figure><p>The short answer is yes, if we let it.</p><p>Almost everyone I meet these days is talking about Artificial Intelligence. Most tech vendors say it’s a core part of their offering, but only a few are really using it. There is clearly a gap between marketing fluff and the truth. Which is nothing new.</p><p>Artificial Intelligence is really big news. It is already changing our world, and has potential to go much, much further. Will it cure all disease and solve big global problems, like hunger and poverty? Or will it make us all redundant then kill us off <a href="https://twitter.com/hashtag/skynet">#skynet</a>?</p><p><strong>What is Artificial Intelligence?</strong></p><p>Artificial Intelligence (AI for short) is best defined as <a href="https://en.wikipedia.org/wiki/Artificial_intelligence">intelligence exhibited by machines</a>. This can be observed when a computer, or a robot controlled by a computer, acts in a way which appears intelligent.</p><p>Ants are intelligent to a certain extent. So, perhaps a better definition is <a href="http://www.huffingtonpost.co.uk/andy-hood/why-we-must-embrace-not-f_b_12249014.html">computer systems able to perform tasks requiring human intelligence</a>.</p><p>A computer telling you 2+2=4 is not AI. However, a <a href="https://www.google.com/selfdrivingcar/">self driving car</a> is a good example of a computer utilising as much, if not more, intelligence than a human.</p><p><strong>What’s the difference between machine learning and deep learning?</strong></p><p>Machine learning is related to AI. It is a technique whereby computers (or machines) are given <a href="https://en.wikipedia.org/wiki/Machine_learning">the ability to learn without being explicitly programmed</a>. Rather than follow a set of pre-defined instructions, the machine is trained using large volumes of data along with some clever algorithms, so over time it gets better and better.</p><p>A real world example of machine learning is <a href="https://matthewearl.github.io/2016/05/06/cnn-anpr/">number plate recognition</a>. The ‘machine’ is shown thousands of images of cars with number plates and it then learns to work out which portion of the image contains the number plate and to decipher the number plate into a text string.</p><p>Deep learning is an extension of machine learning. Where machine learning relies on nonlinear or parallel processing (i.e. multiple calculations being performed at the same time, rather than in a linear sequence), <a href="https://blogs.nvidia.com/blog/2016/07/29/whats-difference-artificial-intelligence-machine-learning-deep-learning-ai/http://beebom.com/examples-of-artificial-intelligence/">deep learning requires a neural network</a>, essentially a complex network much like the human brain, which adds multiple layers and connections.</p><p>In simple terms it is machine learning on steroids, which can enable the processing of more data, more quickly and with less manual input.</p><p><strong>What are the benefits of AI?</strong></p><p>There are many benefits AI can bring to us all. Here are three real-world examples:</p><ul><li>Virtual assistants like Siri, Alexa and Cortana might seem like a gimmick, but they can be incredibly useful, saving time and improving quality of life. Having a smart virtual assistant in your car will make your sat nav easier to use and much safer.</li><li>Product and content recommendations are common on most major ecommerce websites. These help consumers get the right solution and businesses provide a much better service. Netflix use AI extensively to recommend films.</li><li>Many banks now use AI to detect fraud, identifying and stopping funds being illegally taken from your account. This is good for everyone, except the fraudsters.</li></ul><p><strong>So why is AI such big news now?</strong></p><p>It can be argued that AI started with the development of a calculating machine by <a href="https://en.wikipedia.org/wiki/Wilhelm_Schickard">Wilhelm Schickard in 1623</a>. However it was not until 1956 that the field of AI research was formally founded at Dartmouth College.</p><p>Then in 1997, IBM developed a computer called <a href="http://www-03.ibm.com/ibm/history/ibm100/us/en/icons/deepblue/">Deep Blue</a> which became the first computer to defeat a chess grand master, Garry Kasparov. Earlier this year (2016), Google developed a computer system called <a href="https://www.wired.com/2016/01/in-a-huge-breakthrough-googles-ai-beats-a-top-player-at-the-game-of-go/">AlphaGo</a> which beat Fan Hui, Europe’s reigning Go champion. Go is an exponentially harder game to master than chess and even Google’s experts were not expecting such a break through till 2025.</p><p>We are seeing more step changes in the capabilities of AI and we are seeing these earlier than predicted. Things are getting faster.</p><p><strong>What’s driving this change?</strong></p><p>There are three critical components needed for machine learning, deep learning and hence AI:</p><ol><li>Data. The more data the machine is given to learn from, the better (e.g. the more accurate the number plate recognition will be).</li><li>Processing Power: The more powerful the machine, the quicker it can learn and the quicker it can apply these learnings (essential for real time applications).</li><li>Algorithms: Smart/clever algorithms are needed to program machines to learn and also to interpret the results.</li></ol><p>According to <a href="https://en.wikipedia.org/wiki/Moore%27s_law">Moore’s law</a>, computers double in power every two years. Over the past 40 years, this exponential growth has been a key enabler in the development of AI.</p><p>But over the last 5–10 years we have seen something else too. The rapid adoption of social networks, mobile technology and now Internet of Things (IoT), mean that we are creating more data than ever before. According to <a href="https://storageservers.wordpress.com/2016/02/06/how-much-data-is-created-daily/">statistics for 2015</a>, we are creating 2.5 quintillion Bytes of data every day, which would fill 10 million Blu-ray discs. And these discs when stacked on one another, would measure the height of 4 Eiffel Towers.</p><p>With the advent of cloud computing, tech giants like Amazon, Microsoft and Google have enabled very cheap storage for (almost) limitless volumes of data. Even more significantly, they are also now providing access to large arrays of extremely powerful networked computers, which in turn employ <a href="https://en.wikipedia.org/wiki/Graphics_processing_unit">Graphical Processor Units (GPUs)</a>. And it is these GPUs which are ideally suited to machine and deep learning.</p><p>There are also an <a href="https://github.com/showcases/machine-learning">increasing number of platforms (many of them open source)</a> which make it easier and quicker to build AI solutions. A good example of this is <a href="https://www.tensorflow.org/">TensorFlow</a>.</p><p>So the promise is exponentially more data, more processing power and better algorithms, accessible to everyone at an affordable price.</p><p><strong>How can I use AI in my business?</strong></p><p>Here are my top tips:</p><ul><li>If you want a mature ‘enterprise ready’ solution, <a href="http://www.ibm.com/watson/">IBM Watson</a> makes sense. It’s a technology platform that uses natural language processing and machine learning to reveal insights from large amounts of unstructured data.</li><li><a href="https://www.salesforce.com/products/einstein/overview/">Salesforce Einstein</a> is a great option, especially if you are already invested in the Salesforce platform, as it promises to add enormous value to the core offerings of Marketing Cloud, Sales Cloud and Service Cloud.</li><li>If you want to build your own AI platform, then look at <a href="https://www.tensorflow.org/">TensorFlow</a> or <a href="https://aws.amazon.com/machine-learning/">Amazon Machine Learning.</a></li><li><a href="https://allo.google.com/">Google Allo</a> is a great example of how AI can power messaging.</li></ul><p>Of course there are many more providers and partners who can help. Just be aware that quite a few are making empty promises.</p><p><strong>Where is this all going?</strong></p><p>Depending on who you believe, the cleverest machines right now have the cognitive ability of a small mammal, although in many cases they can perform individual tasks (like playing Go) much better than humans.</p><p>The scary thing is that they are getting better so quickly. The <a href="https://en.wikipedia.org/wiki/Technological_singularity">technological singularity</a> is the hypothetical advent of artificial general intelligence (also known as “strong AI”). Such a computer, computer network, or robot would theoretically be capable of recursive self-improvement (redesigning itself), or of designing and building computers or robots better than itself. According to the experts, this will happen some time between 2030 and 2045.</p><p>Eventually a machine will have more intelligence than all humans on the planet combined. When you combine this with the speed of development in other fields, like robotics and nano technology, it all gets rather worrying. Imagine these <a href="https://www.youtube.com/watch?v=rVlhMGQgDkY">robotic dogs</a> becoming our next police force or army.</p><p>The doomsday scenario is that we don’t build in a kill switch and machines start to make big decisions without us. Like manipulating the economy, or starting a war.</p><p>However I am remaining optimistic. AI has so many applications to help people, business and society as a whole. We just need to do two things.</p><ol><li>Agree and build standards</li><li>Tech more kids to code</li></ol><p>The recent <a href="https://www.theguardian.com/technology/2016/sep/28/google-facebook-amazon-ibm-microsoft-partnership-on-ai-tech-firms">partnership on AI</a>, formed by Google, Facebook, Amazon, IBM and Microsoft, is a great start.</p><p>We now need to reform our education system - to focus on Maths and coding. If we can make this fun and get more kids to love it, then we’ve got a good chance.</p><p><strong>How do I learn more?</strong></p><p>If you want to know more, follow any of the links in this post. Whatever you do, don’t bury your head in the sand. Read as much as you can, play with it and if possible building something.</p><p>Us humans need to stay ahead of the machines.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=350a0365eddf" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>