Skip to main content

Stanford Institute for Human-Centered Artificial Intelligence

Stanford HAI —

New tool reveals language models’ political bias

A new tool finds that popular large language models have a decided bias on hot-button topics that may be out of step with popular opinion.

Read More
Stanford HAI —

Can we trust generative search engines?

A generative search engine is supposed to respond to queries using content extracted from top web search hits, but there’s no easy way to know when it’s just making things up.

Read More
Stanford HAI —

Why GPT detectors aren’t a solution to the AI cheating problem

At least seven algorithms promise to expose AI-written prose, but there’s one problem: They’re especially unreliable when the author is not a native English speaker.

Read More
Stanford HAI —

What is a foundation model? An explainer for non-experts

These powerful machine learning algorithms sit at the core of many generative AI tools today.

Read More
Stanford HAI —

Tech increases equity in diabetes care

Continuous glucose monitoring with AI-enabled remote support yields better outcomes for kids across the socioeconomic spectrum.

Read More
Stanford HAI —

Will generative AI make workers more productive?

An AI assistant made new call center employees more productive and empathetic with customers, but offered few benefits for more experienced workers.

Read More
Stanford HAI —

Generative AI model can choreograph dance to any music

A new generative AI model can choreograph human dance animation to match any piece of music.

Read More
Stanford HAI —

A simple algorithm could improve congestion pricing and reduce traffic

An algorithm that adjusts tolls based on the number of cars traveling on a road at a given time could improve congestion pricing and reduce traffic.

Read More
Stanford HAI —

Building a precise assistive-feeding robot that can handle any meal

Stanford researchers are creating novel robotic algorithms to help people with disabilities feed themselves.

Read More
Stanford HAI —

How good is AI?

According to most of the technical performance benchmarks we have today, it’s nearly perfect. That probably means it’s time for new benchmarks, Vanessa Parli argues.

Read More