Technology

Are you human? Researcher says as AI evolves we may need credentials to prove it

One researcher at MIT explains the concept of "personhood credentials" and how we may one day need a system to prove our humanity

NBC Universal, Inc.

In the rapidly evolving landscape of artificial intelligence, you may soon need to have what are known as “personhood credentials” to verify that you’re a person and not a bot, not unlike two-factor authentication. MIT PhD student Nouran Soliman explains.

In a rapidly evolving landscape of artificial intelligence, there's a term you've probably never heard of but soon will: "personhood credentials."

They aim to verify you as a real human online, not a bot, as a way to combat issues like fraud and misinformation. But it raises other questions about oversight.

AI is advancing at a breakneck pace, getting smarter and faster every year. That means there are growing needs for checks on the technology.

Nouran Soliman is a graduate student and researcher at MIT specializing in human-computer interaction. She explains why we as humans may need to start proving our so-called "personhood" and how it could work: "First you show up in person physically for that. So that's something the AI cannot do. And then the other part is you show some verification of your identity like your passport. And then you get issued a credential."

A lawsuit against Hingham Public Schools could shape the way schools handle artificial intelligence in the future.

To put the use of AI-generated content online into perspective, 90% of online content could be generated by AI in 2025. This highlights the need for verifying identities. But it also creates new concerns about privacy.

"We are trying to also think about ways of implementing a system that incorporates personal credentials in a decentralized way. So it's also important not to have the power in one place because that compromises democracy," Soliman said.

We saw it before in the 2024 New Hampshire primary, when fake Joe Biden robocalls urged voters not to vote.

And then, there was the AI-generated political propaganda featuring Taylor Swift's image posted by Donald Trump, prompting Swift to come out and publicly endorse Kamala Harris.

Fears over AI's ability to enable large-scale scams is driving the need to help users distinguish fact from fiction.

"There needs to be a lot of thought about how this is implemented," Soliman said.

AI is here to stay, and as it evolves, so must our human strategies for living with it.

When asked about a timeline for when we could see digital credentials in our everyday lives, Soliman said that she estimates that would be closer to 10 years from now, eventually becoming a standard on par with our current use of two-factor authentication.

Exit mobile version