Hacker News Reader

Comment by dr_dshiv | Hacker News Reader

Proposed: The laws were drafted in the age of narrow AI and have little relevance in the age of general AI (LLMs, etc).

I think these parts are relevant, it will be very nice to get this for LLM and image generators:

> Generative AI, like ChatGPT, will not be classified as high-risk, but will have to comply with transparency requirements and EU copyright law:

> Disclosing that the content was generated by AI

> Designing the model to prevent it from generating illegal content

> Publishing summaries of copyrighted data used for training

This means that now if an article was generated by LLM they are legally required to disclose that.

Those last two are just going to make llms harder to use and act dumber. Annoying but not the end of the world. I have no idea how disclosure can possibly be enforced though. There's just no realistic way to distinguish a reasonably good image generation and a photoshop job.

Critiques:

1. "AI generated content": Is it generated by LLMs if LLMs were used in any way? e.g., getting ideas? If you use a single word? Sentence? Paragraph? If a person has edited things down?

2. "Prevent it from generating illegal content": It is illegal to libel. An autocorrect can generate libel.

3. "summaries of copyrighted data": and if it used GPT4 as the evaluator?

> Disclosing that the content was generated by AI

Hmm, hadn't thought of that angle before. So, like, if one uses a copilot to generate code, does it now have to be watermarked somehow?

Reply on news.ycombinator.com