What we learned at the inaugural Artificial Intelligence for Good summit
Impartial.ai was invited to exhibit at the Inaugural AI for Good summit in Geneva last week, co-hosted by the XPRIZE foundation and the International Telecommunications Union, a specialized agency of the United Nations. The summit was excellent and we are extremely fortunate to have taken part in the opening of such an important and defining conversation for humanity. As delegates of XPRIZE, we attended the event alongside four other teams each using AI to tackle some of the world’s greatest challenges. With the UN stepping up as a platform for discussion and cooperation on AI, this will help raise the profile of the issue and encourage policy makers to sit up and take notice.
Here are some of our takeaways:
We should focus less on the “AI” and more on the “Good”. There was a recognition that we as technologists must do more to remain attentive and focused on the issues and concerns of people everywhere, not just on the problems that we are familiar with. Acknowledging this frees us to think creatively about how we can solve problems using the tools we already have at our disposal; for example, there are lots of APIs that are freely and widely available right now.
People really care about doing the right thing. It’s sometimes tempting to think that technologists are all focused on the technology or on making a quick buck. This conference was a great reminder that this simply isn’t true, and that fundamentally people want to do good. It’s our responsibility as leaders in our respective fields to work hard to improve the incentives and increase accessibility of technology to enable the good to happen.
Collaboration takes time, but it’s worth it. Marcus Shingles, , CEO of X-Prize foundation and previously head of Deloitte Consulting’s Innovation Group, did well to point this out in his remarks. The bigger the change, the greater the need for change management. Some may decry the fact that the Summit was more about setting the scene and defining the terms of discussion rather than agreeing on concrete actions, but the reality is that this is critical. Making sure that key stakeholders are involved, particularly on the global stage, will help when it comes to getting those actions implemented.
As we reflect on the event and on what comes next, three questions spring to mind:
What does “good” mean and for who? Especially in the context of such a global conference there are many different definitions of what is good and what is right. One of the panelists raised the point that some of these solutions and approaches to the problems being solved represented a very western viewpoint on good and who was being helped.
Take self-driving cars for example, the jobs they will displace are mostly western in nature but even within that it’s important to examine some of the assumptions made about them. We got the sense there was an implicit assumption that the normal technology adoption curve would not apply to self-driving cars and that market penetration and demand would be enormous from day 1. But on an urban radio show people were polled about whether they would use a self-driving car and 4/5 people said they would not because they wouldn’t trust the technology.
Another oft-cited example to illustrate the issue is around call center automation. Many call center jobs are offshored, so the impacts of technology deployment in one country will be felt by the workforce of another. What responsibilities should the automator have towards the automated?
As we become more dependent on technology, it’s important that we get the feedback of EVERYONE that will be affected by these changes not just assume that we are all on the same page. In one session, an important question was asked by Patrick Jones of IA Collaborative, “Do people have the right to “opt out” of AI technologies or otherwise not be implicated?”
How can we be responsible in our implementation? As the saying goes, “if the only tool you have is a hammer, it’s tempting to treat everything as if it were a nail.” With technology, such a focus on the solution can be a common pitfall, and can have disastrous unintended consequences.
Take this example from Jamaica, where a new road was built between Kingston to Montego Bay, cutting the journey time by two hours by avoiding the old mountain road that you’d have to drive through previously. Trouble is, a lot of family-run food outlets operated along the old road, catering to travelers on their long journey. These businesses were forced to close.
A key theme of the summit was “exponential thinking” i.e. “10–100x improvements”. It’s good to use such unconstrained thinking when defining a vision or desired outcome but again, with such great aspirations, we need to be extra careful with how we approach the implementation. When people are involved, change must sometimes be slow and considered to stick.
We learned from our AirBnB host that the UN has somewhat shifted the emphasis in its messaging from ‘doing good’ to ‘doing no harm’. It’s an important but subtle distinction in terms of how to approach solving global problems and reaching SDGs, and we should apply this to AI too.
How can we help AI Technologist take a user and change management view of technology solutions? With this common dialogue established, getting people on the same page about what we are talking about and why it’s important is a critical next step. As Marcus Shingles pointed out there was a bit of tension between the focus on technology and the focus on people but we believe that striking the right balance will be critical to the UN being able to unleash AI to solve SDGs.
Bringing this all together, we feel that an overarching theme is one of joining the dots; having a better awareness of what the range of perspectives are that matter, so that we can bring those into the conversation. Only then can we hope to have a better understanding of why, so that we can work those into our decision making and be sure that whatever future we design is one that’s sustainable and inclusive.
Incidentally that’s why we feel so strongly about what we’re doing at impartial.ai; it’s precisely this issue of improving awareness to enable better conversations that we are trying to tackle.
Visit us at www.impartial.ai
