Jennifer King, privacy and data policy fellow at Stanford HAI, talks risk of the chat tool and anticipates its proliferation despite those risks.

Expect a rush of new AI tools in 2023, but likely ones that will hit the market without much thought to the business models or societal impact, says HAI Associate Director Russ Altman.

In this article, Axios reports on an HAI event for journalists which focused in part on the critical role humans play in the development and deployment of AI.

AI’s automatic writing tools are making it easier for students to cheat. Associate Director Rob Reich says tech companies and AI developers must agree to self-regulate.

HAI’s Center for Research on Foundation Models launches Holistic Evaluation of Language Models (HELM), the first benchmarking project aimed at improving the transparency of language models and the broader category of foundation models. 

The HAI fall conference challenged attendees to think about what it means to put humans at the center of AI, rather than just “in the loop.”

Artificial intelligence has the potential to transform corporate decision making but getting workers to actually use the technologies is just as important as making sure the systems work in the first place. HAI Faculty Affiliate Melissa Valentine offers five lessons for success.

HAI Policy Director Russell Wald calls for executive action and more congressional hearings and legislation to address the rapidly escalating challenges of AI as identified in the AI Bill of Rights.

Stanford Digital Economy Lab Director Erik Brynjolfsson explains that the goal of AI that mimics human abilities has led to direct economic competition between people and machines—what he calls The Turing Trap. HAI Co-Director Fei-Fei Li weighs in.

HAI Associate Director Curt Langlotz says “will A.I. replace radiologists?” is “the wrong question.” Instead, he says, “The right answer is: Radiologists who use A.I. will replace radiologists who don’t.”

HAI Policy Director Russell Wald says the new Blueprint for an AI Bill of Rights issued by the White House lacks details or mechanisms for enforcement. 

Foundation models present new opportunities for a wide range of professional applications including law and government. JD/PhD Candidate Peter Henderson explains the “Pile of Law” model and data set he and others at RegLab developed for such uses.

The exciting and impressive use of new, powerful foundation models could cause harm to individuals and society but HAI and CRFM are pushing for the creation of an expert review board to set community norms, share best practices and agree to standardized access rules before foundation models are released. We are, as HAI Associate Director Rob Reich puts it, in a race between “disruption and democracy.”

Erik Brynjolfsson and a team of researchers from Stanford, MIT, Harvard, and LinkedIn recently conducted the largest experimental study to date on the impact of digital job sites on the labor market and found that weaker social connections have a greater beneficial effect on job mobility than stronger ties.

HAI Co-Director Fei-Fei Li is highlighted in this article for her significant contributions to AI including the creation of HAI and how it is furthering understanding of the field.

The rapid advancement of foundation models, as well as the ethical concerns around them, is why one year ago, Stanford HAI founded the Center for Research on Foundation Models (CRFM). Despite progress, there is much work left to do.

HAI Privacy and Data Policy Fellow Jennifer King provides insight into the data used to train large language models, and the risks these foundation models pose to everyone’s privacy.

It’s time for a shared, value-neutral way of talking about AI’s risks and opportunities, with policymaker education like HAI’s recent AI Boot Camp being a good start.