In an op-ed published in Time, CSET's Helen Toner discusses the challenges of understanding and interacting with chatbots powered by large language models, a form of artificial intelligence. 

What exactly are the differences between generative AI, large language models, and foundation models? This post aims to clarify what each of these three terms mean, how they overlap, and how they differ.

Director of Strategy Helen Toner explored China's tech standards in China File.

Problems of AI safety are the subject of increasing interest for engineers and policymakers alike. This brief uses the CSET Map of Science to investigate how research into three areas of AI safety — robustness, interpretability and reward learning — is progressing. It identifies eight research clusters that contain a significant amount of research relating to these three areas and describes trends and key papers for each of them.

In an opinion piece for Foreign Affairs, a team of CSET authors alongside coauthors make the case for the use of privacy-enhancing technologies to protect personal privacy.

This paper is the fourth installment in a series on “AI safety,” an area of machine learning research that aims to identify causes of unintended behavior in machine learning systems and develop tools to ensure these systems work safely and reliably. The first paper in the series, “Key Concepts in AI Safety: An Overview,” outlined three categories of AI safety issues—problems of robustness, assurance, and specification—and the subsequent two papers described problems of robustness and assurance, respectively. This paper introduces specification as a key element in designing modern machine learning systems that operate as intended.

In their op-ed for Scientific American, Husanjot Chahal and Helen Toner argue how small data can assist AI breakthroughs.

Conventional wisdom suggests that cutting-edge artificial intelligence is dependent on large volumes of data. An overemphasis on “big data” ignores the existence—and underestimates the potential—of several AI approaches that do not require massive labeled datasets. This issue brief is a primer on “small data” approaches to AI. It presents exploratory findings on the current and projected progress in scientific research across these approaches, which country leads, and the major sources of funding for this research.

As modern machine learning systems become more widely used, the potential costs of malfunctions grow. This policy brief describes how trends we already see today—both in newly deployed artificial intelligence systems and in older technologies—show how damaging the AI accidents of the future could be. It describes a wide range of hypothetical but realistic scenarios to illustrate the risks of AI accidents and offers concrete policy suggestions to reduce these risks.

This paper is the third installment in a series on “AI safety,” an area of machine learning research that aims to identify causes of unintended behavior in machine learning systems and develop tools to ensure these systems work safely and reliably. The first paper in the series, “Key Concepts in AI Safety: An Overview,” described three categories of AI safety issues: problems of robustness, assurance, and specification. This paper introduces interpretability as a means to enable assurance in modern machine learning systems.

This paper is the second installment in a series on “AI safety,” an area of machine learning research that aims to identify causes of unintended behavior in machine learning systems and develop tools to ensure these systems work safely and reliably. The first paper in the series, “Key Concepts in AI Safety: An Overview,” described three categories of AI safety issues: problems of robustness, assurance, and specification. This paper introduces adversarial examples, a major challenge to robustness in modern machine learning systems.

This paper is the first installment in a series on “AI safety,” an area of machine learning research that aims to identify causes of unintended behavior in machine learning systems and develop tools to ensure these systems work safely and reliably. In it, the authors introduce three categories of AI safety issues: problems of robustness, assurance, and specification. Other papers in this series elaborate on these and further key concepts.

"When it comes to high-stakes settings," write Jacob Steinhardt and CSET's Helen Toner, "machine learning is a risky choice." Their piece in Brookings TechStream offers rules of thumb for assessing risk of failure in ML.

CSET's Helen Toner highlights OpenAI's delayed release of GPT-2 and the increased attention it brought to publication norms in the AI research community in 2019. This piece was featured in In the Shanghai Institute for Science of Science's "AI Governance in 2019" report.

CSET’s Helen Toner, Jeff Ding, and Elsa Kania testified before the U.S.-China Economic and Security Review Commission on U.S.-China Competition in Artificial Intelligence: Policy, Industry, and Strategy.

CSET Director of Strategy Helen Toner testified before the U.S.-China Economic and Security Review Commission at a hearing on "Technology, Trade, and Military-Civil Fusion: China’s Pursuit of Artificial Intelligence, New Materials, and New Energy."