
The AI “deepfake” videos offered in support of plaintiff’s motion for summary judgment in this Superior Court case goes way beyond the innocent AI “hallucinations” more commonly reported. Plaintiff in Mendones v. Cushman & Wakefield, Inc., (Cal. Super. Ct. Alameda Cnty. Sept. 9, 2025 No. 23CV028772) submitted two videos that the trial judge easily concluded were fakes: they are each short video clips 7-10 seconds long that loop over and over again (remember the video loop that temporarily foiled Dennis Hopper in Speed?), with a monotone AI voice playing over it with weird and clearly digitally manipulated lip movement.
Then there were images from a Ring cam. How did the judge figure out they were manipulated? Easy: the background was pixelated black-and-white, while the subject in the foreground was in crisp color.
The court considered money sanctions, evidence and issue sanctions, and even a criminal referral, but in the end concluded the only remedy that fit the bill was terminating sanctions. The court struck the complaint with prejudice.
We have a ways to go figuring out how best to integrate AI into the legal profession. This, clearly, is not the way.