Why (un)published EU AI Act already affects chatGPT quality

It has been more than 1 year since chatGPT(v4) came to market and , surprisingly, there are users complaining of.. The reduced quality of the output.

Yes, in less than 18 months we came from astonishment of the results to annoyance.

If you follow recent reddits or even official openAI posts (https://community.openai.com/t/why-is-chatgpt-getting-from-bad-to-worst/617490),

You will see users complaining that the results from the AI solution are actually becoming worse or the app just saying ‘ I can’t do it’ while before serving great answers.

Why is that?

Many think they are just trying to save bandwidth with massive consumption/usage, testing new versions without telling the users etc.?

Really, nobody knows.

I do have an idea though. Nothing about the tech per se it’s at play here. To me, they are testing and attempting to be as compliant as possible with the new regulations (EU AI Act first of all).?

Yes the better, the more creative the answers and more users are satisfied. However, the risk of hallucinations or hacking the app is also higher.

I would not be surprised if, with the increase of ‘I can’t do it’ answers also the cases of hacking and harmful hallucinations also drastically reduced.

Simply put, @openAI wants to serve users without exposing them to the risk of harmful answers (which will soon be against the EU AI act).?

There is a reason why @Porsche and @Ferrari limit their engines to a certain speed: not because they cannot do it but because of exposure to regulations and possibly bad accidents (and bad press for the brand).

Same reason applies here.?

Without being into force yet, the EU AI Act has already been producing results.

#ai #artificialintelligence #business #innovation #technology