In a recent controversy surrounding a legal document central to a challenge against Minnesota’s law on deep fake technology in elections, a prominent misinformation expert, Jeff Hancock, acknowledged utilizing ChatGPT in the drafting process. Hancock, who heads the Stanford Social Media Lab, admitted that the AI assistance led to errors in citations, raising concerns among critics about the reliability of his affidavit. The case is being contested in federal court by conservative YouTuber Christopher Khols, known as Mr. Reagan, and Minnesota state Representative Mary Franson, who argued that Hancock’s filing contained citations that did not exist, branding the document as “unreliable.”
Hancock’s affidavit was intended to bolster the legal stance on the dangers of deep fake technology and its influence on elections. Following assertions from the opposing legal team about discrepancies in his document, Hancock responded by clarifying his use of ChatGPT specifically for organizing research sources. While he stressed that he did not allow the AI to draft the document itself, he conceded that errors arose during the citation process due to what is described as “hallucinations” inherent in AI tools. This situation has sparked a larger discussion about the implications of AI in legal and academic settings.
In a follow-up statement, Hancock defended the substantive content of his filing, asserting the integrity of his expert opinions regarding the influence of artificial intelligence on misinformation. He confirmed that his written arguments were founded on the latest academic research, underscoring his commitment to the veracity of the claims made in his affidavit. Hancock used resources like Google Scholar alongside GPT-4 to merge his knowledge with new research, but this approach inadvertently led to inaccuracies, including two nonexistent citations and one erroneous author reference.
Although Hancock expressed regret for the missteps, he reiterated that there was no intent to mislead either the court or opposing counsel. He publicly conveyed his sincere apologies for any confusion caused by the errors, emphasizing that they do not detract from the essential points and conclusions he reached in the document. Hancock maintained that the main arguments concerning the dangers of deep fake technology and misinformation remain sound and relevant, regardless of the citation inaccuracies.
This incident underscores the ongoing debate over the use of AI tools in sensitive fields such as legal writing, academia, and research. While artificial intelligence can significantly streamline and enhance the research process, it also poses risks, including the potential for generating misleading or incorrect information, as evidenced by Hancock’s experience. Critics emphasize the necessity for careful validation and oversight when integrating AI in professional contexts to avoid undermining credibility and trustworthiness.
As the federal case progresses, the implications of Hancock’s affidavit and the acknowledged errors remain to be seen. The court’s response to the discrepancies and the overall impact on the legal challenge against Minnesota’s deep fake law will likely be closely watched, as this case may set a precedent for how AI-assisted work is evaluated in future legal disputes. The situation serves as a cautionary tale about navigating the intersection between technology and the legal system in an era marked by escalating concerns over misinformation.