4 ms·
The EU's AI Act is now in force
- ChrisArchitect 2y agoOfficial release: https://ec.europa.eu/commission/presscorner/detail/en/ip_24_4123 https://ec.europa.eu/commission/presscorner/detail/en/ip_24_... (https://news.ycombinator.com/item?id=41127693 https://news.ycombinator.com/item?id=41127693)
- a_bonobo 2y agoThey do not mention that OpenAI lobbied heavily for this act to be watered down. https://time.com/6288245/openai-eu-lobbying-ai-act/ https://time.com/6288245/openai-eu-lobbying-ai-act/ >Still, OpenAI’s lobbying effort appears to have been a success: the final draft of the Act approved by E.U. lawmakers did not contain wording present in earlier drafts suggesting that general purpose AI systems should be considered inherently high risk. Instead, the agreed law called for providers of so-called “foundation models,” or powerful AI systems trained on large quantities of data, to comply with a smaller handful of requirements including preventing the generation of illegal content, disclosing whether a system was trained on copyrighted material, and carrying out risk assessments. OpenAI supported the late introduction of “foundation models” as a separate category in the Act, a company spokesperson told TIME. Given how much we see these 'general' systems in political propaganda (think of all the Twitter far right propaganda unmasked by users replying with 'ignore previous instructions') I'm not sure I agree that these models do not fall under the high risk label. The rules for high-risk systems are understandable, mostly increased human oversight and increased transparency: https://artificialintelligenceact.eu/section/3-2/ https://artificialintelligenceact.eu/section/3-2/ I understand that OpenAI does not want increased transparency in line with being the least transparent AI company out there. It's not like they don't have the money to follow these rules.
- dsign 2y ago> Given how much we see these 'general' systems in political propaganda (think of all the Twitter far right propaganda unmasked by users replying with 'ignore previous instructions') I'm not sure I agree that these models do not fall under the high risk label. IMO, political propaganda is still low-risk; it simply means that citizens are panzying around and believing anything that pops in their feeds, and the correct counter-measure is to ban or regulate social media or to find a mechanism to encourage citizens to make better judgements (say, a tax deduction after passing a maths and modelling exam, a check for number of well-researched books read, or anything better than my 20-seconds-of-consideration rant). I also like the current classification, but I hope that as AI's get better and more dangerous, our regulations will evolve accordingly.
- jamesblonde 2y agoMost provisions have a 2 year transitional period, although some come in effect from 2nd Feb 2025. My take is that this act is an attempt by the EU to build some AI companies. They failed at the cloud - where China massively succeeded. They do not want to make the same mistake again. There are provisions to (1) cause trouble for entrants (certifying LLMs), (2) provide national infra for testing (subsidize).
- rasz 2y ago>failed at the cloud - where China massively succeeded by kicking western companies out "national security laws that prevent non-Chinese companies from owning certain critical infrastructure assets" https://www.datacenterdynamics.com/en/news/aws-sells-assets-in-china-to-comply-with-the-law-will-continue-offering-cloud-services/ https://www.datacenterdynamics.com/en/news/aws-sells-assets-...
- RandomThoughts3 2y agoNo, that’s the traditional EU approach of steering through regulations. They do the same for data privacy (GRPD), green investment and limiting the ghc emission of their economy (Taxonomy, CSRD), and probably a ton of other sectors I don’t know as well. The EU also view it as an instrument of soft power as many countries end up basing their own law on the EU as they are often first mover.
- jamesblonde 2y agoI don't understand - are you agreeing or not? (It's GDPR, btw). I also think greenhouse gases are a global issue, not an EU issue - please leave politics out of it.
- hcfman 2y agoSo can a European company develop software and remain outside of the regulation if its not for sale within Europe ?
- RandomThoughts3 2y agoWhy would they want to ? Despite the strong lobbying from US companies echoed here ("but think about our poor shareholders"), the regulation is fairly raisonnable, generally in the public interest and it’s very much possible to develop software while respecting it.
- dogma1138 2y agoHave you read it? It’s 600 pages of edge cases and contradictions.
- hcfman 2y agoDo understand the concept of risk? The penalties for being wrong about something are huge.
- RandomThoughts3 2y agoHow is the penalty huge? This is Europe not the US. I find it hard to believe that after years of GDPR Americans still fail to understand the way European courts work. The worst that will happen at first is a small fine and you are more likely to be hit with a warning and an explanation of what you should have done. Fines only ratchet up from there if you intentionally stay non compliant. Downvote as much as you want. I am eagerly awaiting all your exemples of insane fines levied on small companies for GDPR non compliance. You will be hard pressed to find some. Large companies with legal teams which are intentionally infringing however are hit as they should be. Strangely they are also the one beating the drum the loudest against these regulations, go figure.
- Nasrudith 2y ago"Don't worry, we won't taze you too hard if you aren't a bad guy." isn't a very good argument especially if you want investment.
- TheAceOfHearts 2y ago> Again, the EU has taken a risk-based approach, with most GPAI developers facing light transparency requirements — though they will need to provide a summary of training data and commit to having policies to ensure they respect copyright rules, among other requirements. What exactly does this mean? Will AI companies allow copyright holders to opt-out of having their data or works used as part of training? Are there tools to prune data from already trained models? I'm not sure how questions related to copyright are going to be resolved with existing AI models.
- arjunlol 2y agoClassic EU
- uconnectlol 2y agoregulation fixes nothing and just makes everything worse in the big picture > no recording in public this is the only one that sounds like good common sense but isn't even regulation it's just law any sort of legal system would enforce. all these "safety" things means ai offerings are all useless like google after 2003 when anything you query like "list of fad diet" is transformed into instead of an answer to that "problematic" question, an answer to "NVM LOL HERE IS A LIST OF HEALTHY DIETS" > can't train on copyright material that's so dumb. this world has no interesting copyright content coming out for 2 decades, it's literally all sellouts. losing to AI (which will happen regardless of whether the training material contains any copyright content) is exactly what they deserve. don't forget that these are also people who sue random civilians for "damages" that only exit in their head. sellout culture no longer even applies to musicians and book authors, it's also the majority of software too, with people just creating software for part of their resume, or some crowdsource shit, or a blog. any software anyone makes in a corporation is just to host an ad or be a dev tool for people who just host an ad
- maximus-decimus 2y agoIf you think copyright in general is dumb, then the U.S. should drop it together. Keeping copyright for humans but allowing "copyright laundering" by using A.I. is dumber IMO.
- rsynnott 2y ago> but isn't even regulation it's just law EU regulations are law. > any sort of legal system would enforce Many countries (the US, for instance) do not have such a blanket ban. For instance, see this nonsense: https://www.independent.co.uk/tech/friend-ai-artificial-intelligence-b2588853.html https://www.independent.co.uk/tech/friend-ai-artificial-inte... - its whole value proposition is indiscriminate public recording.