Attorney General Brian Schwalb joined a bipartisan coalition of 35 attorneys general calling on X and xAI to stop Grok generating nonconsensual sexual images and child sexual abuse material, remove what exists and act against users who made it. Schwalb said X had "not only enabled the mass production of nonconsensual deepfake pornography but encouraged and profited from it."
A British academic told AFP that X users used Grok to alter photos she had posted, at first changing her hairstyle and make-up, later depicting her in lingerie, in a bikini and pregnant, with requests and results appearing on her own account. She called it a "digital kidnapping" of the body and said she found no way to report the pregnancy image on X.
Who did it: HYBRID AI Hallucination Cases database
An Ontario court set aside the dismissal of a client's fee assessment against law firm Fogler, Rubinoff LLP on 23 January 2026 but denied him costs, because his factum, prepared by his daughter, who is not a lawyer, cited at least two cases that did not exist. She conceded they were created by AI; the factum was withdrawn.
Who did it: HYBRID AI Hallucination Cases database
In Future Field Solutions v. Van Norstrand, Judge Deborah K. Chasanow found that a reply brief from Erik Van Norstrand's side answered arguments never made and contained false factual assertions. Opponents suggested it was drafted with AI, citing incorrect case citations. On January 23, 2026, the court declined for now to issue a sanctions show-cause order, saying the matter may be revisited.
Check Point researchers say the North Korea-aligned group KONNI is targeting developers and engineering teams on blockchain and crypto projects in Japan, Australia and India. The group uses PDF lures dressed up as product briefs to deliver a PowerShell backdoor the researchers describe as AI-generated, complete with extensive comments and clean structure. The backdoor gathers hardware details and checks for debugging tools.
Deepal CEO Jiang Hairong said the carmaker will use legal means to preserve and verify evidence and firmly crack down on paid 'water army' trolls. His statement followed a blogger's finding that 'Arcfox T5 vs Deepal S07' articles were flooding multiple platforms with obvious signs of AI generation. The blogger wondered whether a manufacturer had arranged them or AI accounts were chasing traffic.
In a FACE Act case brought by Femhealth USA (Carafem), Chief U.S. District Judge William L. Campbell Jr. struck defendants' arguments built on fabricated statutory language and quotes from nonexistent cases, calling it "troubling" and warning of possible sanctions. Rather than apologise, the defendants said Carafem sought "to distract the Court." The judge granted Carafem $5,000 in statutory damages under the Act.
