Opinion

The recent OpenAI Hugging Face hack has raised questions about the safety of frontier A.I. models. Helen Toner, the executive director of Georgetown’s Center for Security and Emerging Technology, discusses ways to improve A.I. safety without pausing development.

In July, the A.I. company Hugging Face discovered that it was hacked by an OpenAI agent. Helen Toner, a former OpenAI board member, breaks down the details of what happened on “The Ezra Klein Show.”

Why does A.I. do things it wasn’t asked to do? Helen Toner, the executive director of Georgetown’s Center for Security and Emerging Technology, explains on “The Ezra Klein Show.”

The damage was real and long-lasting.

By Bret Stephens

It doesn’t take much for small powers to gain a big advantage over the powerful.

By Thomas L. Friedman

Readers respond to a guest essay about the U.S. failure in Afghanistan and the Taliban’s oppression of women. Also: the use of A.I. in medicine.

In Georgia, Mara Gay sat down with Black voters to hear how they think America is doing — and what they make of the Democratic Party today. “Messy. All over the place,” Jarrett Willis, an entrepreneur, told her. “There’s no structure; it’s nothing to rally behind. It’s lacking that umph factor.”

In Prairieland, prosecutors are using antiterrorism laws we’ve seen before.

By Rozina Ali

The president likes to boast about the “greatest economy we’ve ever had.” His record falls far short.

The Trump family’s crypto business exemplifies all of the sleaziest aspects of the president’s code of conduct.

By Thomas B. Edsall