AI Industry 12 Jun 2025

Fake Law defends Pillow Empire

Hallucinations were so `24 - that they kept happening in `25. If you look up AI fails, it's easy to find company after company that was burned by AI getting creative after it had replaced the human workforce. Some would argue that everything has been fixed. New strategies and design patterns to safety checks and guard rails that promise to make AI the full replacement for an expensive work force, others would point to recent cases that prove the opposite.

AI Placeholder - Chris
Rusty Robot Pillow

Defendant AI Makes up case law

On April 23, federal Judge Nina Y. Wang of the District of Colorado identified "nearly thirty defective citations" in a brief submitted by Mike Lindell's defense team. A few example of the problems found in the AI assisted case law included :

  • Citations to cases that "do not exist"
  • Legal principles attributed to decisions that contain no such language
  • Cases from labeled as from the wrong jurisdiction
  • Misquotes of actual legal authorities
  • Completely fabricated case references (ie "Perkins v. Fed. Fruit & Produce Co., 945 F.3d 1242, 1251 (10th Cir. 2019)"

The problems occur when humans hallucinate a world where the AI is a rational thinking replacement for humans. In a world projected to reach 8.2 billion people in 2025, there are too many different potential outcomes to guess the future based on our past. It was evident last year in the numerous cases of AI gone wild, and the problem simply are not going away while leaders imagine that AI is a quick fix to budgetary woes. I'll talk later about the hardware, cloud and energy costs, but for now, just focusing on the use of the technology, the hype alone is not providing the promised value.

So is the AI rein of terror over?

Not likely. When used with appropriate strategies, there is value in using statistical models as exactly that - predictions based on past behavior. Guided with human oversight, these models can help highlight patterns, brainstorm creative, and even dig up new solutions based on past experiences. There is value in AI, and the open source community has democratized the technology to the point where even small business can find extreme value in reasonable use cases. I just don't personally believe that putting a thinking robot at every point of contact is one of them.