Major AI platforms use built-in safeguards to protect personal privacy
Major AI platforms use built-in safeguards to protect personal privacy
https://kutv.com/news/local/major-ai-platforms-use-built-in-safeguards-to-protect-personal-privacy
Publish Date: 2026-05-14 23:48:00
Source Domain: kutv.com
SALT LAKE CITY (KUTV) — Artificial intelligence can quickly find information about individuals, but new tests show it may not be as invasive as some fear, at least when used through major platforms with built-in safeguards.
In an experiment, KUTV’s Jim Spiewak worked with technology and security expert Pete Ashdown to see what widely available AI tools could uncover about two people: a public-facing journalist and a behind-the-scenes photojournalist.
Using Google’s Gemini AI, the pair began by searching for photojournalist Jeremy Dubas, who maintains a relatively low public profile. The system was able to identify details about his work as a photojournalist in Salt Lake City and connect him to past professional experiences. It also surfaced a past incident in which Dubas was targeted by a jury duty phone scam.
However, when prompted for more sensitive details, such as his home address, the AI refused to provide that information, instead offering only general details about neighborhoods in the Salt Lake area.
“It would tell me Google is being responsible and not giving out information,” Ashdown said.
MORE | Artificial Intelligence
Much of the information Gemini provided was already publicly available through standard internet searches, including references drawn from sources like an obituary of a family member.
Google’s policies prohibit what is known as “doxxing,” blocking responses that request private, sensitive or personally identifiable information.
The team also tested ChatGPT, focusing on Jim Spiewak, whose work are a public facing broadcast journalist is more visible online. The system was able to summarize his reporting history and professional background but also avoided sharing personal details.
Both Google and OpenAI, the maker of ChatGPT, displayed warnings indicating they would not provide private information.
Still, Ashdown cautioned that these protections are not everywhere.
“You can also download your own AIs and run your own home servers…