DS in the Real World

Our world is changing.

We are more connected now than at any other point in history; armed with a wealth of information at our fingertips, the geographical distance that once kept us apart has been erased by the digital age.

This transformation has altered the way we interact, how we do business, and how we stay informed. It has given a voice to those who did not have one and created greater transparency and accountability across our sociopolitical landscape. And yet, through this increased connectedness, we’ve seen how different our views of the world remain.

While the last few years have seen more open conversation about these differences — and the underlying biases that drive them — we have also seen the impacts they can have when these biases are ignored; particularly in the world of technology, where automation is both a blessing and a curse.

Inspiration

I’ve followed the early political career of Alexandria Ocasio-Cortez and have often found myself drawn to her approach and message. As her meteoric ascent continues, AOC’s trademark clear and direct approach to the issues of our time has been a welcome change of pace in an otherwise static space.

There have however, been moments when her stance on less-discussed issues have given me pause. During a talk at this year’s MLK Now event, AOC discussed the role technology plays in our every day lives and made headlines with the following comments:

“Algorithms are still made by human beings, and those algorithms are still pegged to basic human assumptions. They’re just automated assumptions and if you don’t fix the bias, then you are just automating the bias.”

It’s not surprising that this drew criticism from a handful of conservative pundits. However, what struck me was how conflicted I was with my own thoughts on the subject. While the intention of the comments were clear, this blanket statement that (all) algorithms were biased because they’d been created by humans didn’t sit right with me.

Can algorithms be biased? Of course.

Do all algorithms contain bias? Of course not.

Right?

Fascinated by this mental game of ping pong, I decided to dig in further to re-evaluate my own perceptions and learn more from some of the experts leading in this space. What follows is a brief summary of my own research: what algorithms are, how they are constructed, what pre-defined biases may indeed be automated, and steps we as a society can take to address these concerns.

Algorithms 101

Before we can fully understand algorithmic bias, we must first decompose the moving pieces of these algorithms themselves. Generally, the word “algorithm” is used to describe some form of machine learning/artificial intelligence. While there are many techniques for developing such a model, at its highest level there are four basic steps:

1. Collect a set of training data and feed it into a machine
2. Let the machine… well, learn
3. Fine tune any parameters until you have an acceptable level of accuracy
4. Use the resulting formula to predict something, given a set of inputs

Algorithms are used in nearly every digital experience we interact with; they recommend products, tell us which of our friends are in the photo we took, and suggest things we might want to do next.

While these may seem harmless, other uses can be more questionable: calculating the credit risk of a loan application, using facial recognition as a means of identification for law enforcement, or automating resume review and selection while searching for the “most qualified” candidates.

It’s clear that the latter set of use cases are more likely to end up as a headline in a New York Times cover story. Yet in all of these examples, we’ve already seen the negative impact algorithmic bias can have when gone unchecked.

Surely the creators of these products didn’t set out to intentionally introduce these forms of automated bias, right? So what gives?

It turns out that, while we’ve gotten very good at teaching our machines to learn, we’re still fundamentally flawed as human beings when it comes to selecting the underlying data to learn from. Whether it be underrepresentation of a particular demographic or over-indexing on another, we’re still guided by our own preconceived notions of the world around us and it’s often reflected in the data sets we use to develop these algorithms.

”Basic Human Assumptions”

So, we’ve established that even without malicious intent, algorithmic bias is possible given the nature of how machine learning models are constructed and fed data. But why is it so hard to collect data sets representative of our population in the first place? To answer that question, we turn to the study of unconscious bias.

We all have to two types of biases: conscious and unconscious.

According to UCSF’s Office of Diversity & Outreach:

“Unconscious biases are social stereotypes about certain groups of people that individuals form outside their own conscious awareness. Everyone holds unconscious beliefs about various social and identity groups, and these biases stem from one’s tendency to organize social worlds by categorizing.”

It’s important to note that unconscious biases don’t make us bad; they just make us human. However, these stereotypes shape our view of the world, and it’s imperative that we understand these biases before hoping to construct representative data sets for fair algorithms.

The good news? Unconscious biases can be changed over time. To minimize their affect, however, we must be proactive in identifying our own biases and take steps to address them.

Here’s a short checklist to get you started:

1. Take a test, like the IAT, to discover more about your own biases and become more self-aware of how and why we may make certain decisions.
2. Explore many of the free online resources available, such as Grovo’s Unconscious Bias Microlearning lessons, managing bias in the workplace by Google and Facebook, or Microsoft’s interactive eLesson.
3. Make a conscious effort to take time and reflect on decisions being made, particularly when those decisions have impacts on others.

OK, so now we have a basic understanding of how algorithms are constructed. We also know that unconscious bias can play a role in the data fed into our machine learning networks and can take steps to ensure we’re thinking about representation as we collect data in the future.

All set right? Not quite.

Just as important as it is to be self-aware of our own biases, we must also understand the potential bias applied to some of the most popular data sets and algorithms already in use. Recent studies have continued to highlight areas of such concern and, as Joy Buolamwini demonstrates in her fantastic TED Talk, something as ubiquitous these days as facial recognition remains a serious concern when evaluating what constitutes the gold standard benchmark.

Can Algorithmic Bias be Solved?

It was after listening to Joy’s talk that I truly understood the gravity of AOC’s comments. The world we live in is full of bias, developed over thousands of years, and forged into the very fabric of our community. So why shouldn’t we expect the same to exist within our algorithms?

Machine learning is not a new concept, but the exponential growth of computing power has accelerated adoption across the industry and the rise of Artificial Intelligence as a Service has only added fuel to the fire — without the requisite understanding of its consequences. While it may not be possible to rid all bias from our algorithms just yet, there are steps we can take now to minimize their impacts.

So where do we go from here?

As practitioners, self-awareness is an absolute must. By understanding our biases, engaging in active conversations about their affects, and creating stronger audits along the way we can be more thoughtful as new data sets are curated and resulting algorithms are developed. Better diversity and representation across our engineering and data science teams also can have a marked impact — though easier said than done in an industry already known for its poor diversity numbers.

On the data front, IBM has demonstrated that open sourcing large and carefully audited data sets allow teams to focus on advancing algorithmic techniques rather than worrying about the quality and distribution of their training data. However, this is not always practical, as curation of such data can be extremely time consuming and expensive, which many companies claim as a competitive advantage.

Other groups, like Google (What-If Tool) and IBM (AI Fairness 360), are developing sophisticated open-source tools to help teams are proactively detect bias and visualize performance across different social groups. These tech giants may be leading the industry in this field, but they’ve also been in the spotlight for a series of mishaps over the years and have a vested interest in addressing some of the issues they’ve helped create.

While each of these ideas represent steps in the right direction, there’s simply not a one-size-fits-all solution at this point in time. A proactive strategy combines several approaches to ensure all possible bias is considered, measured and removed during the development of a new model. Much like our own unconscious biases can be changed, so too can the bias in our algorithms.

Conclusion

Now more than ever, it is critical we continue to educate and advocate for greater transparency and periodic auditing of the outcomes these algorithms produce. Machine learning will continue to become more ingrained in our daily lives and, as we’ve seen, when gone unchecked it has the potential to cause serious unintended consequences.

As you think more about the ways these algorithms influence your life and those around you, I’d strongly encourage you to click through the links and learn more for yourself. Regardless of your political views, this is an issue we as a society simply cannot dismiss.

I was asked recently to describe my philosophy on leadership. While I’ve long held strong beliefs about what it means to be a great leader, I’ve never formalized my thoughts in a way that would be shareable with others. As I started to reflect on my own experiences I began to realize this was harder to articulate than I’d originally thought. It also became painstakingly clear how important this exercise is for anyone who aspires to lead.

What’s this leadership thing really all about? What separates the truly great leaders from the good ones? At the end of the day, what exactly is it that I stand for and how do I want to bring that into work every day?

Below are the foundations of my leadership approach, highlighting those values most important to me. This shouldn’t be viewed as a comprehensive list. Instead, my hope is that this serves as a conversation starter to reflect on how to build, support and grow teams of your own.

Principle #1: Trust & Empower

If there was one thing I could instill in anyone wanting to lead, it would be to establish a culture of trust from day one. Trust your team to make decisions. Trust they will obsess over your product. Trust that when things go sideways you’ll be at their side to help right the ship. By trusting and empowering our teammates, we create a level of engagement and ownership that is difficult to otherwise obtain.

This is the foundation for everything we do and it’s imperative to realizing a team’s full potential. It’s a shared responsibility by everyone; new hires in all roles should be expected to lead regardless of title or years of experience. This principle is the focal point for building future leaders across the organization.

Undoubtedly, mistakes will be made. It’s in these moments that we must prioritize learning over failure. By reinforcing our belief in each other we demonstrate first hand that failure is acceptable, as long as we’re learning.

Principle #2: Treat People Like People

Resources. Far too often I’ve heard that term used to describe people. We must not lose sight of the human element of leadership which is so critical to building a healthy working environment.

Embracing a diversity of thought, experience and beliefs is something we must strive to get better at every day. We’re all different, but what makes us unique is also what drives us to push the bounds of how we see the world around us. Championing a safe environment where these different views and approaches can be shared is incredibly important. It’s not enough to hire a diverse group of people without also giving them the support they need (and should demand) to feel their thoughts are heard and respected.

In today’s society it’s easier than ever to find ourselves in an echo chamber of like-minded people and ideas. As leaders it is our responsibility to set clear expectations on respect, collaboration and support across the team. There’s sure to be tension when we ask people to challenge their core beliefs, but it’s in that space where some of our most creative and impactful work is born.

Principle #3: Set Your Vision and GO!

Once your team has a sense of trust, empowerment and safety it’s time to dial it up and focus on attacking your hardest problems. To do this, you’ll need to have a clear vision of what you’re working towards. Having a shared north star keeps everyone focused when making decisions and weighing tradeoffs.

Great leaders not only set that vision, but relentlessly push their teams to understand and embrace it. Encourage everyone to take action and make the vision their own. Shout it from the rooftop. Reiterate it every chance you get. Never settle. Stay on the offense.

A safe working environment is only half of the equation. Without passion and belief in what you’re building, it’ll be too easy to sit back and get complacent. Have a startup mentality, regardless of the type of organization you’re in. It’s your job to bring constant positivity and excitement every day; not only to where you’re headed, but also the journey you’re on!

Principle #4: Unleash Human Potential

Your team is really humming now. They feel empowered, they trust each other and they’re excited about the future you’re building together. All done, right? Not even close. Just as important as it is to obsess over your product, you need to obsess over the people around you and investing in your team is non-negotiable.

Continued learning, both structured and unstructured, keeps our skillsets sharp and brings fresh perspective to how we approach the problems we’re trying to solve. As a general rule of thumb, you want to magnify your team’s strengths and help identify & develop their weaknesses. While there is no silver bullet here, what’s most important is finding the right opportunities for each individual that align with their own career goals.

Maintaining a healthy dialog about personal growth, providing frequent, candid feedback and having a vested interest in everyone’s success will continue to strengthen the bond you’ve built. After all, as a leader your team’s success IS your success.

Conclusion

Building teams and nurturing a positive environment is no easy task. What I’ve outlined above is far from a complete strategy, but if you’re able to approach your work with these four tenets in mind, you’ll be well on your way.

Above all else, believe in the power of people. Nothing is more exciting than seeing a group of individuals rally around a common cause to solve a problem together. So much of our ability to lead effectively comes from having a solid team who trust in one another, aren’t afraid to make decisions, and challenge each other to bring their best every day.