Becoming an upstander in AI Ethics
Abhishek Gupta
Founder and Principal Researcher, Montreal AI Ethics Institute | Director, Responsible AI @ BCG | Helping organizations build and scale Responsible AI programs | Research on Augmented Collective Intelligence (ACI)
We don't remember the words of our enemies but the silence of our friends.
- MLK
As we've seen the enormous upheaval in the field of AI ethics over the past 3 months, I think it behooves us to think a little deeply about the role all of us can play in making a meaningful, positive impact on the world. This idea of becoming an upstander in AI ethics is particularly powerful and I believe that in 2021, this is the right way to help create a healthier ecosystem for us all.
A. Why is this important
As I had spoken about in my piece on Why Civic Competence is needed in AI ethics in 2021, I believe that it comes with an additional rider that we need to act on that competence. We routinely come across scenarios where we can raise our voices (respectfully) and point our injustices when we see them happen around us or to us. Sometimes we shy away from taking action due to reasons that concern our safety and of those who we cherish. But, as we see in the opening quote, sometimes the cost of that decision can be onerous.
The following are some of the key drivers for why I think we need to become upstanders:
- Vigilance has benefits
- Avoiding the veil of ignorance
- Passiveness is also a stance
1. Vigilance has benefits
Until we have robust regulations and other safety mechanisms in place, we need to become more vigilant and push the envelope in terms of who is empowered to call out the injustices and when.
An interesting piece of work that came out of Amnesty International talks about having a crowdsourced map of where surveillance technology has been spotted in the city is a great step that epitomizes this approach. This boosts how vigilant we are and helps us become an upstander backed with empirical evidence. This evidence adds more substance to our arguments when we stand up and ask for change to happen.
2. Avoiding the veil of ignorance
Even if you think that some of these things don't affect you at the moment, don't forget that the idea from Rawls on the "veil of ignorance" in which you imagine what the rules and mechanisms of interactions should be in a world if you don't know which side of those interactions you would land on.
Essentially what this means is that even if you're not being discriminated against at the moment, you don't know what will be the case when things evolve in the future. It is better to be an upstander fighting against injustices even when they don't concern you now so that you have someone who has your back when inevitably you might yourself become the target of some other form of injustice that is perpetrated by technological systems.
3. Passiveness is also a stance
Some might argue that staying passive and letting things proceed as they are, hoping that the market will correct itself might be one approach to this. And some also argue that not taking an active stance can help them avoid controversy and stay outside the fray of all these messy issues.
But, remember that a passive stance is still a stance and as the opening quote from MLK shows, this is something that will be noticed by others and will shape their perception of the kind of citizen that you are.
B. How to go about it
Talk is cheap and action is everything!
So, that brings up to the question of how to actually go about doing this in the first place once we agree that taking a more active role and becoming an upstander is an important idea.
The following form a great starting point to go about achieving this goal:
- Show up even when there is little progress
- Find others around you who are interested in the subject as well
- Work with communities first-hand rather than relying on proxies
1. Show up even when there is little progress
A lack of progress is often a big deterrent and discouraging factor when it comes to taking an active role in standing up for the right things. Fatigue from sluggish development and negligence of genuine interests can dishearten even the strongest proponents. Yet, as I had outlined in my article Why Civic Competence is needed in AI ethics in 2021, these moments are especially critical for us to show up and try to make an impact.
In Physics, there is a concept called limiting friction that talks about a critical threshold up to which a constant force needs to be applied to get things moving and after crossing that threshold, it becomes slightly easier to move the object as one moves from the domain of static to kinetic friction.
These moments of little progress bear a lot of similarity to this idea of limiting friction and my advice to those working in the space is to add just an extra little bit of push to get things moving and creating progress in the space when there seems to be none.
2. Find others around you who are interested in the subject as well
There are blossoming communities that are local and contextual that are now talking about the subject of AI ethics. While we don't yet have an effective discovery mechanism for all of them, the journey to becoming an effective upstander in AI ethics will certainly be easier if you have some companions for the road.
The Montreal AI Ethics Institute with its learning communities and public workshops is a great way to meet such like-minded folks across the world. Sharing the work that you're doing with the community as a way of raising interest and inviting them to participate in your endeavours and sharing your time and expertise to further boost their efforts is always a great first step.
3. Work with communities first-hand rather than relying on proxies
I can't stress this point enough!
As I had mentioned in my article To achieve Responsible AI, close the “believability gap”, there are many people who are doing really meaningful work who are unfortunately invisible in the current ecosystem because of the way we view credentials and accolades as a proxy for who we should invite to different conversations.
When it comes to working with communities, from our vast experience doing so at the Montreal AI Ethics Institute, we've found that the most effective way to gain insights is to work directly with the communities on the ground who are facing the issues from injustices of algorithmic systems rather than relying on proxies who may or may not represent their interests accurately.
While the work required to actually go and speak with various communities requires deep efforts and time, it is definitely worth it as we get a more comprehensive understanding of the interests of people while being able to address them in a meaningful manner - hopefully with their participation and support too.
C. What to do?
All of what we have discussed so far offers strong reasons and motivations for us to act but as we've said before, talk is cheap and nothing without action. So what should we do then exactly?
The following actions present a great starting point for anyone looking to make a real difference:
- Educate yourself to understand the nuances
- Find channels through which you can constructively share feedback
- Help to build a community around you if it doesn't exist already
1. Educate yourself to understand the nuances
For an upstander to be effective, one must first and foremost educate oneself to arrive at a nuanced and balanced understanding of the issues so that the arguments put forward in defence of an issue are substantial.
It also helps to direct efforts for remediation in a direction that will yield change that is actually going to make a difference vs. superficial changes that might put off the conversation for a bit without actually addressing the issues at their core.
The advantage of educating yourself then will also help you to counter those superficial measures by pointing out how they might fail given your deeper understanding. Ultimately, a great way to push for higher levels of accountability in the activism that is undertaken by people.
2. Find channels through which you can constructively share feedback
Being an upstander doesn't mean that we have to engage in activism that pushes against constructive approaches. We are looking to create change and that does come with the acknowledgement that we need to work with existing stakeholders and actors so that we can enact change without just bringing up issues and then doing nothing to help those trying to put those recommendations into action.
Channels to provide feedback, both from those bringing up issues to those who are going to be responsible for putting the recommendations into practice and vice versa need to be accessible and transparent so that trust doesn't become an unwitting sacrifice in the exercise of sharing feedback and seeing that feedback be considered and incorporated.
3. Help to build a community around you if it doesn't exist already
As I've had a chance to articulate in many places, notably with the establishment of the Montreal AI Ethics Institute, a community is going to be very important as you embark on this journey to make changes to the world.
It all starts with the kernel of an idea but it requires us to also have the capacity to fight through failures and challenges as they will inevitably arise. Having a community to lean back on will be essential then. There are a large number of regional and domain-specific communities that are working on tackling these issues but if you find that they don't quite address the problems you're seeing, go ahead and start one up! As a bonus, feel free to reach out to the Montreal AI Ethics Institute team and we might even be able to help you there as well.
To read more insights and learn about ethical AI, you can sign up for The AI Ethics Brief by the Montreal AI Ethics Institute.
You can also find out more about my work here - https://atg-abhishek.github.io
Director at IBM leading AI vision for strategic growth
4 年In the day full of distractions, this is the article, I read start to finish in one shot. You post is simply inspiring. "Passiveness is also an instance" should provoke everyone's thoughts. Just signed up for your newsletters and I am going to read more on your work. Kudos !!
Professor of Artificial Intelligence
4 年Hear hear Abishek. The time has come for each of us to consider our roles in the future of #artificialintelligence - as researchers, developers, users, and those with the responsibility for regulation and lawmaking - at the individual, group, organisational, regional, national, and global levels. I strongly believe that the individual level should be upheld and that the group and organisational level isn't just about the "big tech" players - who have a huge responsibility for sure - but is about how innovative smaller companies, and the individual people within them, behave. We all have a role here.
AI & Ethics. Privacy (CIPP/C). Diversity. Speaker. Founder, Ethically Aligned AI I Linked In Instructor I Adjunct Faculty
4 年"there are many people who are doing really meaningful work who are unfortunately invisible in the current ecosystem because of the way we view credentials and accolades as a proxy for who we should invite to different conversations" - Abhishek Gupta - great article, and in particular, that line really resonates with me. That's something I'm thinking about a lot - how to invite more and different voices to be part of these important conversations.