Introduction
In August 2025, xAI sued its former engineer Xeuchen Li for stealing trade secrets relating to Grok and take them to OpenAI. Later on, xAI also sued OpenAI in September 2025 claiming misappropriation of its trade secrets. Recently, OpenAI filed its response to xAI’s claims.
In my earlier posts, I had highlighted that xAI seems to have a reasonably strong claim against OpenAI.
Basic Analysis
There is a Latin maxim, dolosus versatur in generalibus, meaning deceivers deal in generality, and that is exactly what OpenAI seems to be doing in its response to xAI’s allegations relating to misappropriation of trade secrets. Though OpenAI has attempted to deny all the allegations of xAI in a general manner, yet there are some aspects and logical anomalies that warrant attention.
1. Instead of dealing with the specifics, OpenAI is trying to portray Elon Musk in bad light and is trying to play the victim card by hilariously claiming on its website that “Elon is trying everything he can to slow down OpenAI for his personal benefit. Here, we present the facts.”
2. OpenAI is focusing on xAI’s admission that Xuechen Li never joined OpenAI. However, by doing so, it is conveniently brushing aside the fact a legal action has already been brought against Xuechen Li by xAI.
3. Para 36 of OpenAI’s answer is interesting. It seems a like a specific denial. Unless OpenAI has access to the standards that are applied by xAI in its works, it is unfathomable as to how it can claim to know whether xAI secures employee devices or does not adopt proper standards for protecting confidential information. OpenAI must be queried on this.
4. The tonality of generality of other arguments is evident from the fact that according to OpenAI, everything they did was a part of the standard recruiting process and there was nothing fishy in their conduct. OpenAI also seems to claim that the allegations by xAI do not contain the exact details. It ought to be remembered that usually, claimants in such matters involving trade secrets keep crucial evidence close to their chest and reveal the same strategically before the court at an opportune time.
5. Basically, it is a borderline foolish argument by OpenAI claiming that xAI has not revealed its trade secrets. xAI does not need to and even if it will, it will likely do so in a sealed cover meant only for the eyes of the court.
Ending Remarks
I will keep following this case and share my insights from time to time. If you have any specific queries, please feel free to ask.
Also, please feel free to check out my other articles involving AI and law such as How to Train LLMs using Copyrighted Books