How do we successfully govern AI systems that can act autonomously online, making decisions with minimal human oversight? SRI Faculty Affiliate Noam Kolt explores this challenge, highlighting the rise of AI agents, their risks, and the urgent need for transparency, safety testing, and regulatory oversight.

A new study by SRI Graduate Affiliate Michael Colacci sheds light on the frequency of biased outcomes when machine learning algorithms are used in healthcare contexts, advocating for more comprehensive and standardized approaches to evaluating bias in clinical AI.

As AI systems grow more powerful, ensuring their safe development is critical. A recent paper led by David Duvenaud with contributions from Roger Grosse introduces new methods to evaluate AI sabotage risks, providing insights into preventing advanced models from undermining oversight, masking harmful behaviors, or disrupting human decision-making.

SRI researcher Silviu Pitis draws on decision theory to study how the principles of reward design for reinforcement learning agents are formulated. He also aims to understand how large language models make decisions by examining their implicit assumptions. Pitis has received a prestigious OpenAI Superalignment Fast Grant to support his research.

At SRI’s annual academic conference, leading researchers from diverse fields came together to tackle the complexities of AI alignment and how to better understand the social impacts of data-driven technologies. 28 distinguished speakers presented new approaches and ideas to better understand how these technologies are impacting our world.

In We, The Data, Wendy H. Wong argues that we need to reboot human rights to match the pressures of our data-intensive world. Exploring surveillance, facial recognition technologies, data rights, and more, Wong makes a compelling case for why the public needs to engage Big Tech as data stakeholders.

In a recent article in PLOS Digital Health, SRI Graduate Affiliate Vinyas Harish explores how cross-sector partnerships can enable new technologies for more effective healthcare during emergencies, and the importance of robust governance frameworks.

SRI Associate Director Peter Loewen argues that to unlock the value of artificial intelligence for the public sector, governments must double down on the importance of human and democratic values.

ChatGPT’s ability to respond to user prompts in ways that are eerily similar to humans has sparked headlines and concerns around the potential impacts of large language models. In a panel on TVO’s The Agenda, Gillian Hadfield discussed why the system is a “game changer.”

How can we harness the economic benefits of AI innovation? In June 2022, the Schwartz Reisman Institute convened a roundtable with the Business Development Bank of Canada to explore how AI can be better used in financial services, featuring panelists from financial institutions, leading startups, universities, and government. The event’s findings have been issued in a new report that details the importance of responsible AI and new approaches to regulation.

In a session at Absolutely Interdisciplinary 2022, SRI Associate Director Peter Loewen, Jonathan Stray of the Berkeley Center for Human-Compatible AI, and Taylor Owen of McGill University discussed what methods and principles might be used to redesign the algorithms that decide what billions of people see in accordance with democratic values.

The Schwartz Reisman Institute’s academic conference hosted eight panels featuring 30 presenters, with sessions offering innovative responses to the challenges of today’s technological landscape, including questions of data privacy, explainable AI, evolutionary approaches to system design, digital rights, and recommender algorithms.