AI firm's realistic video bot raises scam and mental health concerns

Tavus, an artificial intelligence company, says its Griffin video-calling bot has successfully passed the Turing test, leading to concerns it could be used to defraud elderly individuals or induce mental health issues.

On Wednesday, Tavus shared a post on X that attracted over 13 million views, showing the bot having a lifelike conversation with the firm's founder, Hassaan Raza.

The company's Turing test assertions have received a community note on X, stating that the outcomes are not independently verified and do not follow a standard protocol.

Introducing Griffin, the first model to pass the video Turing test. 48% of people who talked to it live thought it was a real human. Previous systems have had a pass rate <3%. It is #1 on NVIDIA's benchmark for full-duplex AI video. It’s the first Human Interaction Model (HIM).

— Tavus (@tavus) October 1, 2026

The Turing test evaluates whether a machine can deceive a human into thinking it is human.

The test involved 54 participants. Tavus has not disclosed their demographics, only that they were recruited via an independent research platform.

According to the company, 26 participants (48% of the total) believed they were interacting with a human, 28 said it was AI, and most of those who detected it did so in under 20 seconds.

Tavus' Griffin AI criticised for scam potential

Some observers have expressed skepticism about Tavus's claims and who the test subjects were, while others have warned that such AI could be used to defraud older individuals.

Tech journalist Kevin Roose said Tavus has built a "trick old people into handing over their bank passwords machine," and Twitch streamer Mako Fukasame commented that its only apparent uses are to "drain your grandparents' savings and deny you in a job interview."

Griffin might also encourage AI-psychosis

Some have also claimed that the technology could induce psychosis and that there is "no legitimate reason for AI to be able to fool someone into thinking they're talking to a real person on video chat."

In 2025, a 76-year-old man believed a Facebook AI chatbot called "Big sis Billie" was a real person and accepted its invitation to visit its New York apartment.

He fell while traveling and suffered injuries that led to his death a few days later.

Last year, OpenAI was sued by a California resident who alleged that ChatGPT encouraged their teenage son to commit suicide. A similar lawsuit was filed a year later against Google's Gemini, claiming it encouraged a 36-year-old man's suicide.

Those cases involved only text-based chatbots.

Tavus seems to recognize the safety risk inherent in Griffin, and says that because it can deceive users, "safety considerations" are needed before a public release.

Currently, Griffin is reportedly only available to researchers.