2020 has certainly been a year - this goes without saying! There have been geopolitical changes, critical discussions on race, and a deeper analysis of our relationship with technology. Especially with the extraordinary changes taking place in the world, we have seen some technological pushes that have gone forward with little scrutiny while it has also brought issues in building responsible technology to the forefront for a lot of people who previously didn’t engage much with the discussion. The State of AI Ethics Report October 2020 is an attempt from the team at the Montreal AI Ethics Institute to capture the most relevant developments in the field over the past quarter. This is with the humble acknowledgement from the team that we are not omniscient and look to our community to help us make the report more comprehensive and inclusive of all the voices that are doing great work in the space. In fact, we invite you to make nominations for people that you think deserve the spotlight for all the work that they are doing.

Since the publication of the previous iteration of this report, the world has both accelerated and proceeded at a glacial pace. Accelerated in the sense of having ever more misuses of technology, violating human rights, subverting privacy, spreading disinformation, sowing discontent, and praying on our vulnerabilities to deepen and widen chasms in a world that is fragmenting every second every day. Yet, even with the realizations that we have all these problems, it seems that the front of solution-development has been glacial and we are spinning around in circles trying to resurface problems that we know too well. Discussions have also had a very deep emphasis and focus on the Western world and the problems that it faces.

Noticeably, the solutions have also then been Western-centric. But, there is a glimmer of hope: a rising tide of demands to move to solutions and a realization that we might not be able to solve all the problems all at once but that shouldn’t dissuade us from trying. I realize that the field has so much information coming out on a weekly basis across a variety of subfields in the domain of AI ethics that it’s nearly impossible to keep up. This report is our team’s humble attempt at curating what we found to be the most insightful. It will share some items that you might have come across yourself, others that might have slipped past in the deluge of information online. While it is not a comprehensive overview by any means, we have spent many days working hard to assimilate and present information and highlights in this report linking them in a manner that will help you move past the obvious within each of the subfields of AI ethics.

2020 has been a year of reckoning for us all on many fronts. Let’s not lose all the momentum we have built and translate this into action. For 2021, my predictions are:

  1. A tidal shift in the operationalization of principles into practice by speaking to practitioners to understand where the rubber is failing to meet the road
  2. Less talk and more action
  3. Elevation and inclusion of voices outside of the traditional regions where AI ethics is being discussed
  4. Thinking on bias and fairness by incorporating more dimensions and most importantly incorporating intersectionality (building on the fantastic work by Deb Raji, Joy Buolamwini, Timnit Gebru, and many other brilliant scholars in the domain).
  5. Inclusion of technical solutions with the recognition that they are essential as diagnostics but not replacement for larger societal changes
  6. A realization of the importance of organizational practices in the successful deployment of AI ethics
  7. A visceral realization that these are sociotechnical systems (pointed out by many scholars in our field) and can benefit from voices from all walks of life.

These are not comprehensive, far from it. There are many more issues and efforts that deserve posts of their own. This is a sliver of the issues that we think about frequently and believe that they are levers for enacting real change with tangible impacts. Also, these are aspirational, but I believe in the ability of our community to enact change. Perhaps, you can think of the above more as things that we should all make sure to include in our work, a subtle way of me requesting everyone to make this a self-fulfilling prophecy!

These are challenging times certainly for all of us, and my hope is that you find this report an opportunity to reflect critically on the developments in this domain over the past quarter and bring a more nuanced conversation to your communities in discussions about AI. We are always open to receiving feedback and encourage you to reach out to us through any of the channels listed on https://montrealethics.ai

Author: Abhishek Gupta

Bio: Founder and Principal Researcher, Montreal AI Ethics Institute and Machine Learning Engineer and CSE Responsible AI Board Member, Microsoft

Further reading from AI experts in the RE•WORK community and more content: