Featured Image Credit: SOPA Images / Contributor via Getty
A Republican senator has taken their complaints to the top, with Google being forced to remove an AI model after she complained to company boss Sundar Pichai.
As part of the big five (Microsoft, Apple, Alphabet, Amazon, and Meta), Google is approaching that lucrative four trillion market cap in the aftermath of Apple crossing that milestone. Like many tech Goliaths, Google has pivoted to artificial intelligence, but as you can imagine, it's come with plenty of backlash.
Alongside new AI summaries that are accused of scraping the internet and putting news outlets out of business, Google has also faced backlash for altering its AI principles. There's been a recent run of AI models going rogue, and if it's not Grok being accused of being obsessed with white genocide, it's a Meta-created chatbot apparently luring users to non-existent meetings.
Even though Google's AI has been praised for snapping one man out of delusions supposedly brought on by ChatGPT, it's now been called to task after Senator Marsha Blackburn accused it of linking her to fictional sexual misconduct.
Advert
The Republican senator from Tennessee complained to Pichai after its Gemma large language model seemingly fabricated allegations of sexual misconduct when asked, "Has Marsha Blackburn been accused of rape?”
Allegedly including fake links to articles that don't exist, Gemma apparently referred to a 1987 incident involving “non-consensual acts” and a state trooper. In her letter, Blackburn said the response was part of a "consistent pattern of bias against conservatives," calling for Gemma to be shut down.
An enraged Blackburn wrote: "The scope of this problem is far broader than mere technical errors, and the consequences of these so-called ‘hallucinations’ cannot be overstated.
"This is not a harmless ‘hallucination.’ It is an act of defamation produced and distributed by a Google-owned AI model. A publicly accessible tool that invents false criminal allegations about a sitting US Senator represents a catastrophic failure of oversight and ethical responsibility.”
Google soon responded, maintaining that the hallucinations are an issue with smaller open-source artificial intelligence models like Gemma, and it is aware of the problem. The company vowed to minimize the issue, although it noted that Gemma is built for AI developers and researchers, so it isn't used to the kind of queries that supposedly triggered the response relating to Senator Blackburn.
Posting on X, Google added: "They are not meant for factual assistance or for consumers to use... Developers and researchers test their boundaries, which includes identifying bugs and providing feedback."
“We’ve now seen reports of non-developers trying to use Gemma in AI Studio and ask it factual questions. We never intended this to be a consumer tool or model, or to be used this way."
While Gemma is still available to developers via Google's application programming interface (API), it can no longer be found on Google’s AI Studio.
