Google AI ties user's full name to DUI article about someone else, Reddit poster says

After searching their name online, one Pennsylvania resident was stunned to find a Google AI answer tying them to a 2018 DUI charge.

Here's what to know

Their Reddit post described a small button connecting Google's AI response to an article about a DUI charge. However, different first and middle names distinguished the resident from the person in the piece, despite their shared last name and location. 

The use of their full name in the AI answer was central to the poster's complaint. They said they had no history of legal trouble — not even a speeding ticket — yet the AI answer attributed the charges to them.

When a commenter asked whether someone with the same name had a DUI, the poster clarified the distinction.

"I have a pretty uncommon name," the OP said. "As far as I know, there's only one other person with my same first/last name and nobody with my name has DUI charges."

Search results about namesakes did not personally identify the poster, one commenter argued. They cited their own experience sharing a name with several people, including a skateboarder and a federal judge.

Another commenter said the issue was the mismatch between the poster's full name and the name in the linked article.

What can be done?

One commenter suggested consulting a lawyer as a fallback if an initial request to Google to correct the result proved unsuccessful.

A job offer withdrawn by an employer because of the AI-generated DUI allegation was the hypothetical case the commenter used to illustrate concrete harm. In that situation, they said, seeking legal advice would be especially important.

"AI results are often confidently wrong but there's been no major legislation that addresses mistakes," another warned.

Where can I learn more?

While major concerns about AI focus on the technology's extensive energy and water needs, incorrect and potentially damaging responses also fuel irritation and suspicion. 

• A Tennessee woman targeted $10 million in damages after an alleged AI-related wrongful imprisonment.

• Google AI's Gmail and Photos connections could help target advertisements using personal information.

• A Utah mother said Meta AI named her daughter and surfaced her birth details after she shared a video of her.

• In Florida, a woman faced solitary confinement after misidentification involving a Flock camera.

• Anthropic said Claude wrote an invented homicide tip that was submitted to Philadelphia police.

Get TCD's free newsletters for easy tips, smart advice, and a chance to earn $5,000 toward home upgrades. To see more stories like this one, change your Google preferences here.