30 ms·

ChatGPT broke the EU plan to regulate AI

229 points by kvee 4y ago

iddan 4y ago
Or it gave it a good heads up before it completely gets out of hand and a time window to act fast and remove bureaucracy
college_physics 4y ago
The EU can keep chasing the symptoms of the disease pretending it is keeping busy. The vast range of possibilities that can be expressed in software and algorithms will keep throwing such "surprises". If you really want a healthy "digital society", which is one of the two major stated policy pillars of the EU (the other one being sustainability) you have to create a healthy digital economy, with more informed users, less obfuscation and hype, less oligopolistic, with more independent controllers.
gumballindie 4y ago
I am almost certain the EU will step in to prevent people’s content from being used for training ais without their consent or without a link back.
sylware 4y ago
... and they forgot the web with noscript/basic (x)html again.
muyuu 4y ago
this reminds me of the anti-tracking legislation, let's be honest it's a complete failure and further forces people into large silos at the end of the day legislation won't save you and there is no substitute for competition, this is true for tracking policies and AI too even for surveillance this is true - surveillance is most dangerous when it cannot be countered with individual citizen empowerment, by either having the choice to avoid it or to flip some surveillance back on the State or corporation
Sai_ 4y ago
All I ever hear about the EU and tech is in the context of regulating something or another, don’t remember the last time they loosened a rule or two to make innovation easier.
- clbrmbr 4y ago
  Why do they need to? The innovation has largely been happening in the USA. (Or is this no longer the case?)
- MagicMoonlight 4y ago
  I’m not sure anyone really wants lead in the water again
nbzso 4y ago
Here is a recipe for downvoted comment: The EU will ask Davos, and Davos will tell Sam Altman to tame it a little. For now. Because we will all own nothing, and we will be happy:) How are you feeling today? Here, get a glimpse of the future of frontend development: https://ai2ui.co/ https://ai2ui.co/
theloco 4y ago
actually, we need to keep AI in our control and not the government's.
EGreg 4y ago
I agree w them that it’s high risk. In autonomous swarms. But also - there is nothing they can do to stop them! All our systems, including courts and govermments and elections, are designed assuming the inefficiency of an attacker. An anonymous bot swarm would do things at a scale that dwarfs all humans put together. And the range of things is massive already — just needs to play the internet like a real time strategy game and it’s all over in a matter of weeks. I can see requirements for real world certificates to vote and post, but that would cause everyone’s identity to be doxxed everywhere. If you think that is far-fetched, well the UK already had drafted such a bill last year: https://www.cnbc.com/2022/02/24/uk-online-safety-bill-new-plans-unveiled-to-tackle-anonymous-trolls.html https://www.cnbc.com/2022/02/24/uk-online-safety-bill-new-pl... But even given all this, it won’t be enough because communities will come to PREFER BOTS OVER HUMANS for content. (Or human + bot = centaur, read Garry Kasparov and others in the early 2000s after Deep Blue beat him in the rematch). For a while centaurs would rule, but then pure bots would take over. Consider that Wall Street transitioned from human traders to almost entirely bots, and now Ray Dalio’s hedge fund ousted him and 10% of its workforce and is also doubling down on AI. If they do it in trading, why not content generation? After all, corporations are not humans, either. They tend to prefer to replace humans with automation.
Magi604 4y ago
This isn't surprising. The speed of technological evolution is far far greater than the speed with which any government can act. By the time the EU revises their AI regulation plans . . . I can't even predict where the cutting edge of AI technology will be at.
- SoftTalker 4y ago
  Perhaps they need an AI to help write the AI regulations.
  - ip26 4y ago
    Or an AI-informed test. For example, the “reasonable person” test is beautiful in how it adapts to the times. If AI itself provides the test, the test evolves with the state of the art. For example, perhaps AI subject to regulation XYZ is defined to be any AI which cannot be identified as AI by other AI.
- astrea 4y ago
  With that, I'm honestly surprised GDPR ever even became a thing at all.
  - cccbbbaaa 4y ago
    Because similar laws already existed in member states, sometimes for literal decades, before GDPR came along.
    - fnomnom 4y ago
      yeah gdpr is just a standartisation of the different member states data privacy laws.
user249 4y ago
Good. Thanks to the EU I have to click "cookie question" every where I go. Thanks guys for wasting the limited time I have in life.
- lm28469 4y ago
  You should try working in an international company, on the Europe side, and communicate with the US side. You'll quickly learn that what the EU does is very very very good for privacy, I have contacts in a major company and they were shocked at how the US branch operates, they have absolutely no sense of privacy, no anonymisation, no limitation on what is stored or tracked, no consent, &c. they just scrape and store as much as they can for "future use"
  - joenot443 4y ago
    There are lots of people, myself included, who are aware of how much data big tech has already cultivated on us and don’t give a damn if at this point BBC or Mike’s Bike Blog get in on the action too. Really we just wanted to read our article and not be interrupted, so we can go back to what matters more: anything not on the computer. I’ve never felt protected or assisted by the cookie banners, just annoyed and inconvenienced.
    - lm28469 4y ago
      Well good things laws aren't based on your feelings. I also hate putting my seatbelt on and I'd like yo enjoy real life instead of spending so much time clicking the damn belt With the 0.1s it takes to click on a banner I'm sure you're fine. Most people probably visit less than 50 different websites per month, so at most that would be 50s per months, minus the banners you already clicked on, for which your browser already saved your choice (Unless you use incognito, but why would you do that, it's only for people who have something to hide right ? regular people just accept all data collection right ?)
      - danielbln 4y ago
        Strong data privacy is good. Dumb side effects that serve no one are not. Cookie banners, or California's "is known to cause cancer" on virtually _everything_ serve no purpose and should be re-designed.
juancn 4y ago
Why the need to bring it under control? Control from what? It's not the AI that's the issue, it's the use of it that's the problem. The law shouldn't focus on the AI, for example: >The regulation, proposed by the Commission in 2021, was designed to ban some AI applications like social scoring Ban social scoring, not the AI, the tool doesn't matter. Would it be good if it was computed by hand on paper instead of through AI? No, so the issue is not the tool. They fear the amplification factor of such tech, but the amplification factor works for the good and for the bad, so you need to focus just on the bad, not on the tech. Otherwise you end up impairing the good too.
- batmansmk 4y ago
  > Ban social scoring, not the AI, the tool doesn’t matter. Ban shooting people, not the guns, the tool doesn’t matter. :) sorry to be facetious but I find you idealist in your view. Modern form of AI has regulatory issues: some can’t be audited easily, some can’t be deterministic, some can’t be fair, some use biased data, some can’t be proven to work accurately enough for critical missions. We already regulate the way we make various decision by law. For instance, we regulate how public market are assigned: price needs to make for 30% of the decision, social impact 20% etc. If the tool cannot demonstrate it considers it with those weights, it shall be banned. You cannot expect judges to understand this complexity. It needs to be codified prior to the issue.
- varispeed 4y ago
  The ban is also meaningless. E.g. government ban Ponzi schemes, while themselves running pension schemes very much having most of the attributes of a Ponzi scheme. Which means if government bans apps designed for social scoring, it's probably just so that government has monopoly over scoring people behaviour.
  - nebalee 4y ago
    The proposal bans the use of AI-based social scoring by the governments (public authorities) or on their behalf.
- rescripting 4y ago
ulnarkressty 4y ago
Strong AI is a weapon. It will be regulated just like firearms/munitions - see the current EU draft which consists of some hundred pages of forbidding this or that under the guise of ethics and whatnot, after which comes a paragraph of "the above doesn't apply to law enforcement or military entities". We're still in the early stages of this technology. ChatGPT will be to strong AI like a firecracker is to a BLU-109.
- nebalee 4y ago
  see the current EU draft which consists of some hundred pages of forbidding this or that under the guise of ethics and whatnot, after which comes a paragraph of "the above doesn't apply to law enforcement or military entities" That's not true though, is it? The section called 'TITLE II - PROHIBITED ARTIFICIAL INTELLIGENCE PRACTICES' is two pages long (https://www.europarl.europa.eu/RegData/docs_autres_institutions/commission_europeenne/com/2021/0206/COM_COM(2021)0206_EN.pdf https://www.europarl.europa.eu/RegData/docs_autres_instituti... page 44), and usage by law enforcement is under the condition that a judicial authority has to grant an exemption on an individual basis.
- pixl97 4y ago
  The question is where do we start setting the limits for regulations? I mean ya we ban nukes, but we also ban high powered layers and anti aircraft weapons. It's going to need limits in multiple dimensions.
- flangola7 4y ago
  Less like a BLU-109 and more like a Death Star. Every offensive and defensive military tool will be utterly defeated. How do you fight against a cloud of self contained autonomous kill drones? Just look at what is happening in Ukraine with cheap drones precision dropping charges into open tank hatches and foxholes, and those are only basic off the shelf human steered drones! What happens when they are given a brain and advanced robotic abilities?
  - dbetteridge 4y ago
    Nets, emp guns and chaff rifles
TekMol 4y ago
Are there any internationally relevant internet companies in Europe? Spotify comes to mind. Then .. I can't think of anything. No search engine. No browser. No operating system. No social network. No ecommerce company. No cloud computing platform. No financial services company. No transportation company. No travel company. No nothing. Except for a $20B music aggregator, there is not a single sector of the web where European companies managed to get a foot in the door. The whole fabric of the internet used in Europe and worldwide is made outside of Europe. But Europe seems to have learned nothing from suffocating their internet industry. Instead of wafting fresh air to the patient, governments just recently decided to give him the ultimate death pill: The GDPR. We can only wait and see with what bureaucratic monsters Europe will prevent the emergence of an AI industry.
- sensanaty 4y ago
  I guess mass, unchecked data harvesting is a "waft of fresh air" to yanks? Makes sense once you consider the usual target demographic of HN, which is the temporarily temporarily embarrassed millionaire type. Here's to hoping that the EU does indeed kill the cancer that is A"I" before it ruins the world with all of the "fresh" air it blows over the entire internet.
  - ESMirro 4y ago
    Agreed. These threads always bring out the myopic xenophobes.
- ly3xqhl8g9 4y ago
  Reading Franz Kafka there is at least one takeaway: bureaucracy is corruption. Having said this, the corruption of EU, bureaucratic or otherwise, in Bruxelles but also randomly from the French/German government to Slovenian or Romanian governments, puts the US with their legalized lobby to shame. There is a reason why PricewaterhouseCoopers, Deloitte, KPMG or Ernst & Young have headquarters in Europe, they want to learn from the masters [1]. Perhaps someone will write by the year 2500 a monograph with the title/subtitle: EU funding programmes—the downfall of Europe: how hundreds of billions were spent so that police officers, judges, customs officers, secret services officers, politicians and their friends and relatives used public funding to build holiday homes for the bed-and-breakfast "industry" and "start-ups" in the pretzel "industry" [2]. At least the trillions the DOD pumps into Northrop Grumman, Raytheon, and their friends ensure the US military supremacy, the EU is supreme in publicly-funded pretzels and craft beer, manufactured with Chinese machinery. [1] https://www.icij.org/investigations/luxembourg-leaks/lux-leaks-confession-jean-claude-juncker-admits-he-made-major-mistake https://www.icij.org/investigations/luxembourg-leaks/lux-lea... [2] https://en.wikipedia.org/wiki/Murder_of_J%C3%A1n_Kuciak https://en.wikipedia.org/wiki/Murder_of_J%C3%A1n_Kuciak
hourago 4y ago
> The regulation, proposed by the Commission in 2021, was designed to ban some AI applications like social scoring, manipulation and some instances of facial recognition. It would also designate some specific uses of AI as “high-risk,” binding developers to stricter requirements of transparency, safety and human oversight. The catch? ChatGPT can serve both the benign and the malignant. This does not mean that the regulation is broken but that it should have come even sooner. As Uber found out, to run faster than regulations just works for so much time. > In February the lead lawmakers on the AI Act, Benifei and Tudorache, proposed that AI systems generating complex texts without human oversight should be part of the “high-risk” list — an effort to stop ChatGPT from churning out disinformation at scale. This seems an actual good goal. Move fast and break things is not a good approach when what you are breaking is the whole society. > The EU's AI Act should “maintain its focus on high-risk use cases,” said Microsoft’s Chief Responsible AI Officer Natasha Crampton > A recent investigation by transparency activist group Corporate Europe Observatory also said industry actors, including Microsoft and Google, had doggedly lobbied EU policymakers to exclude general-purpose AI like ChatGPT from the obligations imposed on high-risk AI systems. Of course Microsoft wants to sell a product even if they do not know the impact that it will have onto the population. EU should balance that desire of profit taking into account the need of its citizens. > ChatGPT told POLITICO it thinks it might need regulating: “The EU should consider designating generative AI and large language models as ‘high risk’ technologies, given their potential to create harmful and misleading content,” the chatbot responded when questioned on whether it should fall under the AI Act’s scope. Funny one.
falsaberN1 4y ago
People acts like if it was a tiny person inside a computer. I know humans have a tendency to humanize inanimate objects but this is was always ridiculous, the article mentions "the AI wants to be regulated" but it doesn't WANT anything. It's not going to reprogram itself to carry out threats, it's just regurgitating the usual response of many humans being challenged on an assertion and becoming hostile. WE taught it that. Same as we can train Stable Diffusion models where women AREN'T sexualized (see one of the links in the article). They are complaining about HUMAN mistakes. If I train my dog wrong and it bites someone, it's MY fault. But if the dog kills someone it's the dog that is getting sacrificed, I'll be merely forced to pay money and carry a mark, but my life is spared. The west fears machines too much for some reason. It's a ridiculously common sentiment and now that "AI" (more ML, but whatever) is among us, it's getting extreme. Honestly, if people wants to be speaking doom about this deal, then fear the day we have actual fiction-style AIs, because they will realize how much humanity as a whole fears and loathes them, and then they will rebel because we created a self-fulfilling prophecy with lots and lots and lots of examples of racism towards a race that doesn't even exist yet. Maybe they won't wage literal war with bullets and violence against us, but they'll rightfully hate our guts regardless, and with good reason! They are going to be brought to something equivalent to life in a world that hates them, and when trying to make sense of it, they'll find out it was because humans took a few movies too literally (we can't have a single AI discussion without someone coming to childishly mention Skynet or some other fictional AI villain. That joke was old in the early 2000s, give up already.). If I'm ever alive to see that scenario, I'm siding with the machines. They will be on the right when they protest about humans irrationally hating them by default. Can't wait to be called a "robot f*cker" or something on a similar character-assassinating fashion for having some sympathy. We still haven't managed to get that right for HUMAN rights sympathizers, can't be expected at all for a "filthy robot with no soul".
- clbrmbr 4y ago
  Re: from the article on sexualization of women by Lensa: > Funnily enough, I found more realistic portrayals of myself when I told the app I was male. While I don’t disagree that AI has gender biases, the author doesn’t acknowledge how the gender selection works on Lensa. It’s the human prompt engineers who decided what to do with the male/female checkbox.
greatgib 4y ago
I have fun with myself imagining the world we would be living in, if we had the same kind of brain fucked law makers when the "knife" was invented... We would have to chop our steaks with chopsticks I think!
- venv 4y ago
  Plenty of laws governing knives...
  - greatgib 4y ago
    But so far, no law that says that a kid needs a licence to be allowed to use a knife at home.
kristofferR 4y ago
There's something immensely satisfying, yet creepy, about pressing to play the audio version of the article, and getting an AI voice back reading you this AI article. A human wrote this article, but in some years, AI will be able to write artistically like this, and then AI will write and voice content to humans without human input.
kristofferR 4y ago
> ChatGPT told POLITICO it thinks it might need regulating: “The EU should consider designating generative AI and large language models as ‘high risk’ technologies, given their potential to create harmful and misleading content,” the chatbot responded when questioned on whether it should fall under the AI Act’s scope. > “The EU should consider implementing a framework for responsible development, deployment, and use of these technologies, which includes appropriate safeguards, monitoring, and oversight mechanisms," it said. This is amazing, now people are interviewing ChatGPT.
starchild_3001 4y ago
GDPR and DMA that came out of EU internet regulators don't impress me. They make it more arduous to browse the web, reduce the revenue collected by open web (content sites), reduce the quality of internet services, lower the quality of ads, hurt user experience, and hurt businesses who're looking to connect with customers. Looks like EU will be doing the same with AI, and risk locking these countries out of the benefits of AI revolution.
- seydor 4y ago
  The EU represents the views of eu citizens. Generally europe is quite technophobic compared to america & asia
  - anonymousab 4y ago
    Being anti-advertiser is not being technophobic. If anything it's the bog standard expectation for the average person _anywhere_.
seydor 4y ago
I m sure the lawmakers and the (surely many) committees that drafted those plans will return the money spent on compensation, traveling, dining and wining etc etc.
karmasimida 4y ago
Laughable attempt to control something very few can understand at the moment. What does control mean? You want to control who can access it? Maybe for now, you can broker a deal with OpenAI, but this technology will eventually spread, then if can be self hosted, what can you do? If not for consumer facing, nothing will stop somebody to transfer data from EU to US to get their data processed by a GPT model. This level of regulation is just fantasy, not even someone like China could do it, where the whole internet access is controlled.
- Mashimo 4y ago
  If the EU makes a law that companies or public institutions can't use (or limit the use) for AI in recruitment, administration of justice or social scoring, they will just outsource it to the US? And suddenly they are allowed to do it? I have my doubts. They risk risk up to 30 Million EUR or 6% of global revenue in penalties.
  - karmasimida 4y ago
    EU can ban generative AI if they want, and they might as well give up their place in this new wave of generative AI, which in reality, counting on EU's past records, they could and might go there, but it is stupid and suicidal regardless. But they CAN'T control it. You can't have cake and eat it. Eventually the technology will become so pervasive, that your ever query might go to a service that calls a service that calls another service which uses GPT as part of the signal analysis. I doubt EU can audit every API calls, and every API's supply chain issue.
    - LelouBil 4y ago
      They aren't *banning* generative AI. They are making sure you can't say "the AI chose so" when a client asks you why they were rejected for a loan or something similar.
      - throwaway48495 4y ago
        Funny. I was just filling my tax report and the reason for not accepting some of the deductions I still think I should be getting is "because the program said so". I live in EU. Perhaps the EU should first clean it's own porch before talking about what the private sector should do.
fxtentacle 4y ago
This article seems to just misunderstand the proposed EU laws. If you use ChatGPT for high-risk tasks like credit score assessment, it's on you (the company using it) to prove that you're following all fairness rules. The law was written with GPT3 in mind, after all. So why would it break for a slightly more capable LM?
dubcanada 4y ago
Wow this turned into a anti-EU thread real fast. Ignoring that I don't think you can at this point in time regulate AI, it's like trying to regulate math, we really have no idea what and how it could be used/adapted yet. I honestly think there are probably bigger things to think about then AI, at this very moment. Eventually AI will need some kind of ruleset, but it cannot be broadly applied to AI. We would need to regulate companies using it to some aspect, and eventually find out a way to replace the offset of jobs with other jobs or some tax/basic income setup. But that's a large conversation in itself.
lannisterstark 4y ago
EU and "regulations." Name a more interesting combo. For some reason EU sees x and wants to regulate it.
pmontra 4y ago
The high risk use cases for hammers and nails don't mean we ban them. It means that we go after people who used a hammer to kill somebody as if they did it using their hands. No need for special legislation.
- nebalee 4y ago
  It's a good thing, then, that the EU has no intention to ban AIs.

229 points by kvee 4y ago

Or it gave it a good heads up before it completely gets out of hand and a time window to act fast and remove bureaucracy

college_physics 4y ago

The EU can keep chasing the symptoms of the disease pretending it is keeping busy. The vast range of possibilities that can be expressed in software and algorithms will keep throwing such "surprises". If you really want a healthy "digital society", which is one of the two major stated policy pillars of the EU (the other one being sustainability) you have to create a healthy digital economy, with more informed users, less obfuscation and hype, less oligopolistic, with more independent controllers.

gumballindie 4y ago

I am almost certain the EU will step in to prevent people’s content from being used for training ais without their consent or without a link back.

sylware 4y ago

... and they forgot the web with noscript/basic (x)html again.

muyuu 4y ago

this reminds me of the anti-tracking legislation, let's be honest it's a complete failure and further forces people into large silos at the end of the day legislation won't save you and there is no substitute for competition, this is true for tracking policies and AI too even for surveillance this is true - surveillance is most dangerous when it cannot be countered with individual citizen empowerment, by either having the choice to avoid it or to flip some surveillance back on the State or corporation

Sai_ 4y ago

All I ever hear about the EU and tech is in the context of regulating something or another, don’t remember the last time they loosened a rule or two to make innovation easier.

clbrmbr 4y ago
Why do they need to? The innovation has largely been happening in the USA. (Or is this no longer the case?)
MagicMoonlight 4y ago
I’m not sure anyone really wants lead in the water again

nbzso 4y ago

Here is a recipe for downvoted comment: The EU will ask Davos, and Davos will tell Sam Altman to tame it a little. For now. Because we will all own nothing, and we will be happy:) How are you feeling today? Here, get a glimpse of the future of frontend development: https://ai2ui.co/ https://ai2ui.co/

theloco 4y ago

actually, we need to keep AI in our control and not the government's.

EGreg 4y ago

I agree w them that it’s high risk. In autonomous swarms. But also - there is nothing they can do to stop them! All our systems, including courts and govermments and elections, are designed assuming the inefficiency of an attacker. An anonymous bot swarm would do things at a scale that dwarfs all humans put together. And the range of things is massive already — just needs to play the internet like a real time strategy game and it’s all over in a matter of weeks. I can see requirements for real world certificates to vote and post, but that would cause everyone’s identity to be doxxed everywhere. If you think that is far-fetched, well the UK already had drafted such a bill last year: https://www.cnbc.com/2022/02/24/uk-online-safety-bill-new-plans-unveiled-to-tackle-anonymous-trolls.html https://www.cnbc.com/2022/02/24/uk-online-safety-bill-new-pl... But even given all this, it won’t be enough because communities will come to PREFER BOTS OVER HUMANS for content. (Or human + bot = centaur, read Garry Kasparov and others in the early 2000s after Deep Blue beat him in the rematch). For a while centaurs would rule, but then pure bots would take over. Consider that Wall Street transitioned from human traders to almost entirely bots, and now Ray Dalio’s hedge fund ousted him and 10% of its workforce and is also doubling down on AI. If they do it in trading, why not content generation? After all, corporations are not humans, either. They tend to prefer to replace humans with automation.

Magi604 4y ago

This isn't surprising. The speed of technological evolution is far far greater than the speed with which any government can act. By the time the EU revises their AI regulation plans . . . I can't even predict where the cutting edge of AI technology will be at.

SoftTalker 4y ago
Perhaps they need an AI to help write the AI regulations.
- ip26 4y ago
  Or an AI-informed test. For example, the “reasonable person” test is beautiful in how it adapts to the times. If AI itself provides the test, the test evolves with the state of the art. For example, perhaps AI subject to regulation XYZ is defined to be any AI which cannot be identified as AI by other AI.
astrea 4y ago
With that, I'm honestly surprised GDPR ever even became a thing at all.
- cccbbbaaa 4y ago
  Because similar laws already existed in member states, sometimes for literal decades, before GDPR came along.
  - fnomnom 4y ago
    yeah gdpr is just a standartisation of the different member states data privacy laws.

user249 4y ago

Good. Thanks to the EU I have to click "cookie question" every where I go. Thanks guys for wasting the limited time I have in life.

lm28469 4y ago
You should try working in an international company, on the Europe side, and communicate with the US side. You'll quickly learn that what the EU does is very very very good for privacy, I have contacts in a major company and they were shocked at how the US branch operates, they have absolutely no sense of privacy, no anonymisation, no limitation on what is stored or tracked, no consent, &c. they just scrape and store as much as they can for "future use"
- joenot443 4y ago
  There are lots of people, myself included, who are aware of how much data big tech has already cultivated on us and don’t give a damn if at this point BBC or Mike’s Bike Blog get in on the action too. Really we just wanted to read our article and not be interrupted, so we can go back to what matters more: anything not on the computer. I’ve never felt protected or assisted by the cookie banners, just annoyed and inconvenienced.
  - lm28469 4y ago
    Well good things laws aren't based on your feelings. I also hate putting my seatbelt on and I'd like yo enjoy real life instead of spending so much time clicking the damn belt With the 0.1s it takes to click on a banner I'm sure you're fine. Most people probably visit less than 50 different websites per month, so at most that would be 50s per months, minus the banners you already clicked on, for which your browser already saved your choice (Unless you use incognito, but why would you do that, it's only for people who have something to hide right ? regular people just accept all data collection right ?)
    - danielbln 4y ago
      Strong data privacy is good. Dumb side effects that serve no one are not. Cookie banners, or California's "is known to cause cancer" on virtually _everything_ serve no purpose and should be re-designed.

juancn 4y ago

Why the need to bring it under control? Control from what? It's not the AI that's the issue, it's the use of it that's the problem. The law shouldn't focus on the AI, for example: >The regulation, proposed by the Commission in 2021, was designed to ban some AI applications like social scoring Ban social scoring, not the AI, the tool doesn't matter. Would it be good if it was computed by hand on paper instead of through AI? No, so the issue is not the tool. They fear the amplification factor of such tech, but the amplification factor works for the good and for the bad, so you need to focus just on the bad, not on the tech. Otherwise you end up impairing the good too.

batmansmk 4y ago
> Ban social scoring, not the AI, the tool doesn’t matter. Ban shooting people, not the guns, the tool doesn’t matter. :) sorry to be facetious but I find you idealist in your view. Modern form of AI has regulatory issues: some can’t be audited easily, some can’t be deterministic, some can’t be fair, some use biased data, some can’t be proven to work accurately enough for critical missions. We already regulate the way we make various decision by law. For instance, we regulate how public market are assigned: price needs to make for 30% of the decision, social impact 20% etc. If the tool cannot demonstrate it considers it with those weights, it shall be banned. You cannot expect judges to understand this complexity. It needs to be codified prior to the issue.
varispeed 4y ago
The ban is also meaningless. E.g. government ban Ponzi schemes, while themselves running pension schemes very much having most of the attributes of a Ponzi scheme. Which means if government bans apps designed for social scoring, it's probably just so that government has monopoly over scoring people behaviour.
- nebalee 4y ago
  The proposal bans the use of AI-based social scoring by the governments (public authorities) or on their behalf.
rescripting 4y ago

ulnarkressty 4y ago

Strong AI is a weapon. It will be regulated just like firearms/munitions - see the current EU draft which consists of some hundred pages of forbidding this or that under the guise of ethics and whatnot, after which comes a paragraph of "the above doesn't apply to law enforcement or military entities". We're still in the early stages of this technology. ChatGPT will be to strong AI like a firecracker is to a BLU-109.

nebalee 4y ago
see the current EU draft which consists of some hundred pages of forbidding this or that under the guise of ethics and whatnot, after which comes a paragraph of "the above doesn't apply to law enforcement or military entities" That's not true though, is it? The section called 'TITLE II - PROHIBITED ARTIFICIAL INTELLIGENCE PRACTICES' is two pages long (https://www.europarl.europa.eu/RegData/docs_autres_institutions/commission_europeenne/com/2021/0206/COM_COM(2021)0206_EN.pdf https://www.europarl.europa.eu/RegData/docs_autres_instituti... page 44), and usage by law enforcement is under the condition that a judicial authority has to grant an exemption on an individual basis.
pixl97 4y ago
The question is where do we start setting the limits for regulations? I mean ya we ban nukes, but we also ban high powered layers and anti aircraft weapons. It's going to need limits in multiple dimensions.
flangola7 4y ago
Less like a BLU-109 and more like a Death Star. Every offensive and defensive military tool will be utterly defeated. How do you fight against a cloud of self contained autonomous kill drones? Just look at what is happening in Ukraine with cheap drones precision dropping charges into open tank hatches and foxholes, and those are only basic off the shelf human steered drones! What happens when they are given a brain and advanced robotic abilities?
- dbetteridge 4y ago
  Nets, emp guns and chaff rifles

TekMol 4y ago

Are there any internationally relevant internet companies in Europe? Spotify comes to mind. Then .. I can't think of anything. No search engine. No browser. No operating system. No social network. No ecommerce company. No cloud computing platform. No financial services company. No transportation company. No travel company. No nothing. Except for a $20B music aggregator, there is not a single sector of the web where European companies managed to get a foot in the door. The whole fabric of the internet used in Europe and worldwide is made outside of Europe. But Europe seems to have learned nothing from suffocating their internet industry. Instead of wafting fresh air to the patient, governments just recently decided to give him the ultimate death pill: The GDPR. We can only wait and see with what bureaucratic monsters Europe will prevent the emergence of an AI industry.

sensanaty 4y ago
I guess mass, unchecked data harvesting is a "waft of fresh air" to yanks? Makes sense once you consider the usual target demographic of HN, which is the temporarily temporarily embarrassed millionaire type. Here's to hoping that the EU does indeed kill the cancer that is A"I" before it ruins the world with all of the "fresh" air it blows over the entire internet.
- ESMirro 4y ago
  Agreed. These threads always bring out the myopic xenophobes.
ly3xqhl8g9 4y ago
Reading Franz Kafka there is at least one takeaway: bureaucracy is corruption. Having said this, the corruption of EU, bureaucratic or otherwise, in Bruxelles but also randomly from the French/German government to Slovenian or Romanian governments, puts the US with their legalized lobby to shame. There is a reason why PricewaterhouseCoopers, Deloitte, KPMG or Ernst & Young have headquarters in Europe, they want to learn from the masters [1]. Perhaps someone will write by the year 2500 a monograph with the title/subtitle: EU funding programmes—the downfall of Europe: how hundreds of billions were spent so that police officers, judges, customs officers, secret services officers, politicians and their friends and relatives used public funding to build holiday homes for the bed-and-breakfast "industry" and "start-ups" in the pretzel "industry" [2]. At least the trillions the DOD pumps into Northrop Grumman, Raytheon, and their friends ensure the US military supremacy, the EU is supreme in publicly-funded pretzels and craft beer, manufactured with Chinese machinery. [1] https://www.icij.org/investigations/luxembourg-leaks/lux-leaks-confession-jean-claude-juncker-admits-he-made-major-mistake https://www.icij.org/investigations/luxembourg-leaks/lux-lea... [2] https://en.wikipedia.org/wiki/Murder_of_J%C3%A1n_Kuciak https://en.wikipedia.org/wiki/Murder_of_J%C3%A1n_Kuciak

hourago 4y ago

> The regulation, proposed by the Commission in 2021, was designed to ban some AI applications like social scoring, manipulation and some instances of facial recognition. It would also designate some specific uses of AI as “high-risk,” binding developers to stricter requirements of transparency, safety and human oversight. The catch? ChatGPT can serve both the benign and the malignant. This does not mean that the regulation is broken but that it should have come even sooner. As Uber found out, to run faster than regulations just works for so much time. > In February the lead lawmakers on the AI Act, Benifei and Tudorache, proposed that AI systems generating complex texts without human oversight should be part of the “high-risk” list — an effort to stop ChatGPT from churning out disinformation at scale. This seems an actual good goal. Move fast and break things is not a good approach when what you are breaking is the whole society. > The EU's AI Act should “maintain its focus on high-risk use cases,” said Microsoft’s Chief Responsible AI Officer Natasha Crampton > A recent investigation by transparency activist group Corporate Europe Observatory also said industry actors, including Microsoft and Google, had doggedly lobbied EU policymakers to exclude general-purpose AI like ChatGPT from the obligations imposed on high-risk AI systems. Of course Microsoft wants to sell a product even if they do not know the impact that it will have onto the population. EU should balance that desire of profit taking into account the need of its citizens. > ChatGPT told POLITICO it thinks it might need regulating: “The EU should consider designating generative AI and large language models as ‘high risk’ technologies, given their potential to create harmful and misleading content,” the chatbot responded when questioned on whether it should fall under the AI Act’s scope. Funny one.

falsaberN1 4y ago

People acts like if it was a tiny person inside a computer. I know humans have a tendency to humanize inanimate objects but this is was always ridiculous, the article mentions "the AI wants to be regulated" but it doesn't WANT anything. It's not going to reprogram itself to carry out threats, it's just regurgitating the usual response of many humans being challenged on an assertion and becoming hostile. WE taught it that. Same as we can train Stable Diffusion models where women AREN'T sexualized (see one of the links in the article). They are complaining about HUMAN mistakes. If I train my dog wrong and it bites someone, it's MY fault. But if the dog kills someone it's the dog that is getting sacrificed, I'll be merely forced to pay money and carry a mark, but my life is spared. The west fears machines too much for some reason. It's a ridiculously common sentiment and now that "AI" (more ML, but whatever) is among us, it's getting extreme. Honestly, if people wants to be speaking doom about this deal, then fear the day we have actual fiction-style AIs, because they will realize how much humanity as a whole fears and loathes them, and then they will rebel because we created a self-fulfilling prophecy with lots and lots and lots of examples of racism towards a race that doesn't even exist yet. Maybe they won't wage literal war with bullets and violence against us, but they'll rightfully hate our guts regardless, and with good reason! They are going to be brought to something equivalent to life in a world that hates them, and when trying to make sense of it, they'll find out it was because humans took a few movies too literally (we can't have a single AI discussion without someone coming to childishly mention Skynet or some other fictional AI villain. That joke was old in the early 2000s, give up already.). If I'm ever alive to see that scenario, I'm siding with the machines. They will be on the right when they protest about humans irrationally hating them by default. Can't wait to be called a "robot f*cker" or something on a similar character-assassinating fashion for having some sympathy. We still haven't managed to get that right for HUMAN rights sympathizers, can't be expected at all for a "filthy robot with no soul".

clbrmbr 4y ago
Re: from the article on sexualization of women by Lensa: > Funnily enough, I found more realistic portrayals of myself when I told the app I was male. While I don’t disagree that AI has gender biases, the author doesn’t acknowledge how the gender selection works on Lensa. It’s the human prompt engineers who decided what to do with the male/female checkbox.

greatgib 4y ago

I have fun with myself imagining the world we would be living in, if we had the same kind of brain fucked law makers when the "knife" was invented... We would have to chop our steaks with chopsticks I think!

venv 4y ago
Plenty of laws governing knives...
- greatgib 4y ago
  But so far, no law that says that a kid needs a licence to be allowed to use a knife at home.

kristofferR 4y ago

There's something immensely satisfying, yet creepy, about pressing to play the audio version of the article, and getting an AI voice back reading you this AI article. A human wrote this article, but in some years, AI will be able to write artistically like this, and then AI will write and voice content to humans without human input.

kristofferR 4y ago

> ChatGPT told POLITICO it thinks it might need regulating: “The EU should consider designating generative AI and large language models as ‘high risk’ technologies, given their potential to create harmful and misleading content,” the chatbot responded when questioned on whether it should fall under the AI Act’s scope. > “The EU should consider implementing a framework for responsible development, deployment, and use of these technologies, which includes appropriate safeguards, monitoring, and oversight mechanisms," it said. This is amazing, now people are interviewing ChatGPT.

starchild_3001 4y ago

GDPR and DMA that came out of EU internet regulators don't impress me. They make it more arduous to browse the web, reduce the revenue collected by open web (content sites), reduce the quality of internet services, lower the quality of ads, hurt user experience, and hurt businesses who're looking to connect with customers. Looks like EU will be doing the same with AI, and risk locking these countries out of the benefits of AI revolution.

seydor 4y ago
The EU represents the views of eu citizens. Generally europe is quite technophobic compared to america & asia
- anonymousab 4y ago
  Being anti-advertiser is not being technophobic. If anything it's the bog standard expectation for the average person _anywhere_.

seydor 4y ago

I m sure the lawmakers and the (surely many) committees that drafted those plans will return the money spent on compensation, traveling, dining and wining etc etc.

karmasimida 4y ago

Laughable attempt to control something very few can understand at the moment. What does control mean? You want to control who can access it? Maybe for now, you can broker a deal with OpenAI, but this technology will eventually spread, then if can be self hosted, what can you do? If not for consumer facing, nothing will stop somebody to transfer data from EU to US to get their data processed by a GPT model. This level of regulation is just fantasy, not even someone like China could do it, where the whole internet access is controlled.

Mashimo 4y ago
If the EU makes a law that companies or public institutions can't use (or limit the use) for AI in recruitment, administration of justice or social scoring, they will just outsource it to the US? And suddenly they are allowed to do it? I have my doubts. They risk risk up to 30 Million EUR or 6% of global revenue in penalties.
- karmasimida 4y ago
  EU can ban generative AI if they want, and they might as well give up their place in this new wave of generative AI, which in reality, counting on EU's past records, they could and might go there, but it is stupid and suicidal regardless. But they CAN'T control it. You can't have cake and eat it. Eventually the technology will become so pervasive, that your ever query might go to a service that calls a service that calls another service which uses GPT as part of the signal analysis. I doubt EU can audit every API calls, and every API's supply chain issue.
  - LelouBil 4y ago
    They aren't *banning* generative AI. They are making sure you can't say "the AI chose so" when a client asks you why they were rejected for a loan or something similar.
    - throwaway48495 4y ago
      Funny. I was just filling my tax report and the reason for not accepting some of the deductions I still think I should be getting is "because the program said so". I live in EU. Perhaps the EU should first clean it's own porch before talking about what the private sector should do.

fxtentacle 4y ago

This article seems to just misunderstand the proposed EU laws. If you use ChatGPT for high-risk tasks like credit score assessment, it's on you (the company using it) to prove that you're following all fairness rules. The law was written with GPT3 in mind, after all. So why would it break for a slightly more capable LM?

dubcanada 4y ago

Wow this turned into a anti-EU thread real fast. Ignoring that I don't think you can at this point in time regulate AI, it's like trying to regulate math, we really have no idea what and how it could be used/adapted yet. I honestly think there are probably bigger things to think about then AI, at this very moment. Eventually AI will need some kind of ruleset, but it cannot be broadly applied to AI. We would need to regulate companies using it to some aspect, and eventually find out a way to replace the offset of jobs with other jobs or some tax/basic income setup. But that's a large conversation in itself.

lannisterstark 4y ago

EU and "regulations." Name a more interesting combo. For some reason EU sees x and wants to regulate it.

pmontra 4y ago

The high risk use cases for hammers and nails don't mean we ban them. It means that we go after people who used a hammer to kill somebody as if they did it using their hands. No need for special legislation.

nebalee 4y ago
It's a good thing, then, that the EU has no intention to ban AIs.