alt Hacker News> Apple argues that when trade secret information is fed into an AI agent or model that learns from it, that learning “may create irreversible and continually propagating uses of the trade secret.”
This is somewhat of a high impact argument to test. I wonder if the case will eventually get to working this point out.
Replies
A related argument could be whether the output of an LLM that has been trained on GPL code should also be considered to be under a GPL licence.
the fact that llms can spit out books they've been trained on near verbatim lends some credence to this.
Not sure why people are latching on to this. I'm 95% sure the AI agent was not "trained" on this intellectual property but rather the IP was just analyzed and it's usage and details were recorded to the AI agents memory (typically markdown/JSON summary of previous sessions).
The defense IMO is just trying to muddy the waters between "irreversible AI training that updates weights" and "AI learning by just storing secrets in text files".
Even in the 5% chance this AI model actually was fine-tuned or fully-trained (i.e the weights were updated), the employee did this while employed at Apple and used Apple IP as the training data. Even if the output is under fair-use (because it's transformative), (1) any IP created while employed during work duties almost always belongs to your employer and (2) the act of accessing Apple's IP would be (trade secret or regular) theft if done outside normal work duties.
There's already some precedent when dealing with humans reverse engineering systems. We use "cleanroom" approaches where the ones doing the reverse engineering are segregated from those doing the later design and development work. Instead of producing a design document based on the thing being studied, the reverse engineers produce a specification from which a new design, untainted by knowledge of the original design, is produced.
People using ML to try and reverse-engineer and create a "clean" version of things will likely need to use a similar approach. You can't ask one LLM to take in a circuit design as input and produce a specification and design within the same context. The resulting design will be at least partially informed by knowledge of the original design. The way to do this safely (potentially still with suits happening, but safer at least) will be to have one execution to produce a specification, and a second fresh execution taking the specification to produce a design. At least then you know your LLM was not aware of the original design.
Of course, LLM agents "cheat", so you'll also want to be careful to ensure a clean environment if you're using an agent that does not provide access to the original design material.