By Salmana Ahmed

In pursuit of fair and accountable AI

Why We Invested: The AI Now Institute

“How will artificial intelligence become part of our lives?”

That was the guiding question asked by Kate Crawford at the AI Now 2017 Symposium held at MIT Media Lab last year. The resounding conclusion from the day’s discussions? That it already is.

Whether it is to make recommendations about which products we should buy, to determine which stories appear in our newsfeeds, or to identify the likelihood of someone committing a future crime, AI systems are rapidly being integrated into both the private and public sectors, and playing a role in decisions that significantly impact our lives. As the capabilities of these technologies grow with increasing speed, so too do concerns over their unintended consequences, accountability, and relationship to our existing social, economic, and political landscape.

While concerns about AI have increasingly dominated headlines over the past couple of years, the focus of popular debate has largely centered on the more sensational scenarios of artificial general intelligence, stoking fears of an “AI apocalypse”. However, alongside these there has also been a growing call for a lens of inquiry that focuses more on the immediate risks of AI — addressing exactly how and to what extent AI systems are impacting us today. And, while those questions are difficult to resolve, efforts in pursuit of the answers are proliferating, with the AI Now Institute at NYU as one of the organizations leading the charge.

Launched publicly in November 2017, as an interdisciplinary research center based at New York University, the AI Now Institute is dedicated to understanding the social implications of AI. Omidyar Network is proud to be able to support the organization’s important work with core funding of $1.5M over three years, helping to build out its research in four core domains — Labor and Automation, Bias and Inclusion, Safety and Critical Infrastructure, and Rights and Liberties.

The AI Now Institute aims to produce top tier research and to develop robust measurement methodologies and improved development processes for AI across key areas of application. Its work will be a unique contribution to a field that is currently lacking a critical understanding of how to measure and assess the complex and varied impacts of AI systems. It will also focus on building communities of practice, working across its extensive network to develop talent, host regular events and convenings, and work in partnership with other organizations to support, amplify, and share knowledge across the field. AI Now has already begun to establish key partnerships, the first of which falls under its Rights and Liberties research stream and sees it working closely with the American Civil Liberties Union to “map effects of advanced computation on civil rights.”

AI Now’s 2017 Symposium was not the first public offering from the group. In 2016, Kate Crawford and Meredith Whittaker, founders of AI Now, co-chaired the inaugural AI Now Symposium spearheaded by the White House Office of Science and Technology, the National Economic Council, and the Information Law Institute at NYU. The Symposium was a huge success — the discussions and outputs from the event had significant influence on the White House’s own public reports on AI and helped to begin to reframe the mainstream narrative to be more focused on the near-term risks of AI systems. Since then, Kate and Meredith have continued to develop their thinking with input from a range of experts around the most critical social questions related to AI that will define our immediate future. Their most recent report provides an in-depth review of the key issues related to the four domains AI Now’s research will focus on, and culminates in a set of recommendations that includes calls for an end to black box predictive systems in core public institutions like criminal justice and outlines specific approaches needed to address bias in AI and related technologies.

Kate and Meredith bring extensive experience researching and building innovative approaches to complex technical challenges across industry, academia, and civil society. Kate, one of the leading scholars on the social implications of AI, is a Principal Researcher at Microsoft Research and a Distinguished Research Professor at NYU. Meredith, a Distinguished Research Scientist at NYU, founded Google’s Open Research Group, where she has spent over a decade working on issues of measurement, privacy, and security. They have garnered support for AI Now’s mission from a diverse set of organizations and individuals that cut across disciplines and sectors — an effort which speaks to their belief in the critical importance of having diversity of minds working on the issues that the field is grappling with.

At Omidyar Network, we continue to grow our portfolio of investments in, and relationships with, organizations working on issues related to the ethics and governance of AI. Our grant to AI Now comes alongside our recent membership to the Partnership of AI which gives us another opportunity to sit alongside and support civil society organizations, to help advance AI in the public interest.

As the field continues to tackle the challenges related to AI systems, leadership from AI Now will be crucial in making progress on key issues, and ensuring that stakeholders, beyond those in the tech industry, have influence over the decisions made in relation to the development, use, and governance of AI.

The debate around the impacts of AI technologies is growing in momentum, but it is still in its early stages. Many questions remain unanswered, and the complexities of these issues continue to develop. With that in consideration, we are pleased to be able to support the AI Now Institute at a time when its work has never been more important.