Stanford 'misinformation' expert hit by humiliating AI blunder

A Stanford adept connected misinformation has admitted to utilizing AI to fabricate grounds successful a national tribunal case.

Professor Jeff Hancock, a starring adept connected AI-driven accusation and the laminitis of the Stanford Social Media Lab, was brought successful by the Minnesota Attorney General Keith Ellison to support a authorities instrumentality that criminalizes election-related deepfakes. However, Hancock's ain adept declaration, which was partially generated by ChatGPT, contained fabricated information.

Plaintiffs, including blimpish contented creator Christopher Kohls ("Mr. Regan") and Republican Minnesota Rep. Mary Franson, are challenging the 2023 Minnesota law, amended successful 2024, arsenic an unconstitutional regularisation connected speech. Kohls, known for parody videos utilizing dependable cloning to mimic Vice President Kamala Harris, filed the lawsuit.

The plaintiffs' lawyers pointed retired a circumstantial notation to a nonexistent survey by authors Huang, Zhang, and Wang, successful Hancock's declaration.

Suspecting helium relied connected AI to constitute the 12-page document, they called for the full papers to beryllium dismissed owed to imaginable wide inaccuracies.

Hancock aboriginal confirmed the beingness of 2 much AI-generated 'hallucinations,' which manifest arsenic misinformation successful substance and ocular absurdities for generated images.

The AI fabrications extended beyond existing studies. It besides invented a 2023 nonfiction by De keersmaecker & Roets and attributed 4 nonexistent authors to different works.

In an effort to found his expertise, Hancock highlighted his co-authorship of a foundational insubstantial connected AI-mediated connection and his extended probe connected the intelligence interaction of misinformation.

'I person published extensively connected misinformation successful particular, including the intelligence dynamics of misinformation, its prevalence, and imaginable solutions and interventions,' Hancock wrote.

Hancock employed ChatGPT 4.0 to assistance successful his research, instructing the AI to make world citations for circumstantial points. However, the AI instrumentality inadvertently produced mendacious citations and 4 made up 'incorrect' authors

'The effect from GPT-4.0, then, was to make a citation, which is wherever I judge the hallucinated citations came from,' helium wrote.

The plaintiffs exposed Hancock's perjury, arsenic helium had falsely sworn nether oath to the accuracy of his cited sources. Still, Hancock insisted that his errors '[did] not interaction immoderate of the technological grounds oregon opinions.'

A proceeding is scheduled for December 17 to find the destiny of Hancock's declaration, portion Stanford University remains soundless connected imaginable disciplinary action. 

Hancock's lawsuit is not an isolated incident. In February, New York lawyer Jae Lee faced disciplinary enactment for citing a fabricated lawsuit successful a aesculapian malpractice lawsuit, which was generated by ChatGPT. 

Lee was referred to the grievance sheet of the 2nd US Circuit Court of Appeals connected Tuesday aft she cited a fabricated lawsuit astir a Queens doc botching an termination successful an entreaty to revive her client's lawsuit.

The lawsuit was subsequently dismissed.