Technology Companies Have Been Saying This for Years — Now It's on the Record

We have been advising technology companies about the risks of using AI tools for sensitive tasks; it is useful to finally see a case that highlights these concerns. 

In this case between two rival startup app makers, allegations that the departing engineer extensively queried Claude to find out how to misappropriate confidential information and solicit employees without violating a non-solicit obligation highlight how AI prompts and outputs (i) may belong to your employer; and (ii) may become evidence in a lawsuit. 

This case also illustrates how checking a departing employee's last entries with a company-subscribed AI tool is of high probative value and should become part of a company's checklist for departing employees.  

This also brings up the interesting question of how to check for a departing employee's use of public-facing AI tools because the employer may not have direct access to the prompts and responses there. Instead, the former employer may need to send a document retention hold demand to the departing employee, which includes any use of public-facing AI tools. 

As the use of AI tools continues to gain in popularity, I would expect we will see more and more cases filed that contain interesting tidbits of evidence drawn from a person's use of AI tools.

Members-only hospitality platform Dorsia alleged Tuesday in New York state court that its former senior engineer used a company-paid Anthropic Claude account to strategize his departure to join rival startup Rumor and ignored the chatbot's warnings that he'd be "toast" if he brought Dorsia's confidential information with him.

www.law360.com/...