US, UK, EU Sign 'Legally Binding' AI Treaty 51
The United States, United Kingdom and European Union have signed the first "legally binding" international AI treaty on Thursday, the Council of Europe human rights organization said. Called the AI Convention, the treaty promotes responsible innovation and addresses the risks AI may pose. Reuters reports: The AI Convention mainly focuses on the protection of human rights of people affected by AI systems and is separate from the EU AI Act, which entered into force last month. The EU's AI Act entails comprehensive regulations on the development, deployment, and use of AI systems within the EU internal market. The Council of Europe, founded in 1949, is an international organization distinct from the EU with a mandate to safeguard human rights; 46 countries are members, including all the 27 EU member states. An ad hoc committee in 2019 started examining the feasibility of an AI framework convention and a Committee on Artificial Intelligence was formed in 2022 which drafted and negotiated the text. The signatories can choose to adopt or maintain legislative, administrative or other measures to give effect to the provisions.
Francesca Fanucci, a legal expert at ECNL (European Center for Not-for-Profit Law Stichting) who contributed to the treaty's drafting process alongside other civil society groups, told Reuters the agreement had been "watered down" into a broad set of principles. "The formulation of principles and obligations in this convention is so overbroad and fraught with caveats that it raises serious questions about their legal certainty and effective enforceability," she said. Fanucci highlighted exemptions on AI systems used for national security purposes, and limited scrutiny of private companies versus the public sector, as flaws. "This double standard is disappointing," she added.
Francesca Fanucci, a legal expert at ECNL (European Center for Not-for-Profit Law Stichting) who contributed to the treaty's drafting process alongside other civil society groups, told Reuters the agreement had been "watered down" into a broad set of principles. "The formulation of principles and obligations in this convention is so overbroad and fraught with caveats that it raises serious questions about their legal certainty and effective enforceability," she said. Fanucci highlighted exemptions on AI systems used for national security purposes, and limited scrutiny of private companies versus the public sector, as flaws. "This double standard is disappointing," she added.
How ignorant! (Score:5, Informative)
No, it is *not* legally binding.
It does not become binding without US Senate ratification. (not to mention other countries).
Yes, treaties get signed before ratification.
But to call it "binding" in a headline at that stage is hogwash.
hawk
Re:How ignorant! (Score:4, Interesting)
They are referring to the intention of bringing legally binding provisions, not about its current enforcement status. They have been using this wording "the first legally binding AI treaty" for some time. The website, and the pdf brochure which is dated from May (so before it was signed), says "The Convention on Artificial Intelligence [...] is the first international legally binding instrument" https://www.coe.int/en/web/art... [coe.int]
Re: (Score:2)
Once someone shows a "non legally binding" treaty, that could, possibly, mean something.
Until then, the choices are the gamut of misleading, sloppiness, self aggrandizement, and so forth.
Or it can join "self propelled automobile", "seafaring warship", and "repetitive redundancy" . . .
Re: (Score:2)
"legally-binding treaty" is certainly redundant, and the folks at the COE are aware of it to the point they organize workshops on the distinction between treaties and non-legally binding agreements https://www.coe.int/en/web/cah... [coe.int] I think their intention is to reinforce the idea. It is sort of market-speak. They want to write "legally binding" because previous agreements were not binding, so they want the reader to understand the difference. They don't want to use the word "instrument" because it isn't uni
Re: (Score:2)
And THIS agreement, as far as the US is concerned, is NOT legally binding, so to say so is misleading.
As a previous post stated, unless the senate ratifies this, it is in no way legally binding upon the US whatsoever.
Nope, hogwash. (Score:3)
They are referring to the intention of bringing legally binding provisions, not about its current enforcement status. They have been using this wording "the first legally binding AI treaty" for some time. The website, and the pdf brochure which is dated from May (so before it was signed), says "The Convention on Artificial Intelligence [...] is the first international legally binding instrument" https://www.coe.int/en/web/art... [coe.int]
Nope, have to agree with the parent here.
Correct usage is "would be" the first legally binding instrument, not "is" the first legally binding instrument.
It's hogwash, plain and simple.
Re: (Score:2)
Nobody talks like you are suggesting. When a law is being discussed, lawmakers discuss it in the present tense. Using the conditional all the time would make their discussion insufferable.
Re: (Score:2)
"The signatories can choose to adopt or maintain legislative, administrative or other measures to give effect to the provisions.
Francesca Fanucci, a legal expert at ECNL (European Center for Not-for-Profit Law Stichting) who contributed to the treaty's drafting process alongside other civil society groups, told Reuters the agreement had been "watered down" into a broad set of principles.
"The formulation of principles and obligations in this convention is so overbroad and fraught with
Re: (Score:2)
I assume you are right that they failed in creating strong provisions but I think this is a different debate.
by hawk ( 1151 ) on 2024-09-06 1:10 (#64766838)
No, it is *not* legally binding.
It does not become binding without US Senate ratification. (not to mention other countries).
by Okian Warrior ( 537106 ) on 2024-09-06 1:32 (#64766886)
Correct usage is "would be" the first legally binding instrument, not "is" the first legally binding instrument.
The other posters are criticizing the usage of word "legally b
Re: (Score:2)
Re: (Score:2)
So it still has a chance to become legally binding in 56 other jurisdictions in the world who have expressed their intention to sign. Between yesterday and today, there were 10 signatories (including the 3 mentioned in the slashdot headline). Full list: https://www.coe.int/en/web/con... [coe.int]
Re: (Score:2)
You must find the news insufferable, then. From a news search for "new bill proposed":
Drivers who stop in Philly's bike lanes would face fines under proposed bill [phillyvoice.com]
Additionally, the bill proposed Thursday would allow the EPA to manage the rivers [....] [10news.com]
Legacy admissions at Brown would end under this proposed bill [browndailyherald.com]
Online purchases could cost more under proposed bill [wsfa.com]
And so on. People do, in fact, use the potential voice (or sometimes future tense) when talking about bills that are not yet law.
Re: (Score:3)
So you're saying that if Trump wins and Putin wants him to sign it, then there's a chance it will be ratified? Noted!
Nope, in the U.S., our President's signature on a treaty holds no power nor commitment whatsoever. To be ratified, our Senate has to vote for it by a two-thirds supermajority (which is uncommon, because our Senate is usually pretty close to being balanced 50:50 between two political parties that make it a point to oppose each other on nearly every issue). Unlike most other pieces of legislation, our 2nd legislative house (The House of Representatives) has no say in treaties, nor does a ratified treaty need
Re: (Score:2)
And besides...it appears that Putin has endorsed Harris for US President [msn.com]....so...
Re: (Score:2)
Re:How ignorant! (Score:4, Informative)
It's a little more complicated than that. The signatories have committed to writing the rules into their local laws. Obviously that is a democratic process and may fail, but if one of them doesn't do it there may be enforcement by the others.
For example, say the UK parliament refused to pass it into law, the EU could put restrictions on AI developed in the UK, services based in the the UK that may use AI, export of data to the UK that could be used to train AI etc.
Re: (Score:3)
It does not become binding without US Senate ratification.
In fact, most US treaties are not Senate-ratified, but they are binding. There are other approaches available and while Senate ratification is used sometimes, it's not terribly common.
There are basically three different ways the US handles international agreements (what everyone else calls "treaties"):
1. Sole-executive agreements. These are agreements signed by the executive branch (the president, basically), without any congressional action. This can be done if the terms of the agreement bind the U
Re: (Score:2)
The article mentioned the US, UK, and EU have signed it. So, it could still be legally binding for one of the other 3 parties.
Legally Biding (Score:3)
Re: (Score:3)
So when someone breaks the terms of agreement, they'll go to jail?
No, they'll get more venture capital
Re: (Score:2)
So when someone breaks the terms of agreement, they'll go to jail?
Depends on how the treaty terms are enacted in federal law. But "legally binding" doesn't necessarily mean that there are criminal consequences for violating the law, it means that courts can and will enforce compliance. Refusing to abide by a court order could result in jail time for criminal contempt of court, of course, but that would be for violating the order, not for breaking the law. Some laws specify civil or criminal penalties for violation, but that's probably not the case here.
pinkie swear (Score:2)
We pinky swear to keep helmets on our muppets.
Human right to employment. (Score:2)
The AI Convention mainly focuses on the protection of human rights of people affected by AI..
Great! So I take it you’re going to start by giving every human their lost-to-AI-speculation job back? No? Then stop bullshitting about your desire to protect human rights. Being employed is a human right, until governments stop colluding with Greed N. Corruption and define whatever UBI will be to replace human employment.
And it fucking better be an improvement on the current welfare system. Otherwise Greed won’t get a chance to remove that right from humans. Greed and Governments are goin
Re:Human right to employment. (Score:5, Interesting)
My suspicion is, if AI does what it claims to do and completely automates the entirety of the clerical middle class's jobs, then we're gonna need more than a good welfare system, we're gonna need a complete restructure of the economy, because 100 million unemployed people who are used to living in middle class comfort is a recipe for absolute chaos. Complete carnage.
You know how the Zs keep threatening guillotines? That sort of chaos manifested into reality.
I have my doubts however that either of the major parties are capable of summoning the courage to make the changes needed. Chaos might be the reality.
Re: (Score:2)
My suspicion is, if AI does what it claims to do and completely automates the entirety of the clerical middle class's jobs, then we're gonna need more than a good welfare system, we're gonna need a complete restructure of the economy, because 100 million unemployed people who are used to living in middle class comfort is a recipe for absolute chaos. Complete carnage.
You know how the Zs keep threatening guillotines? That sort of chaos manifested into reality.
I have my doubts however that either of the major parties are capable of summoning the courage to make the changes needed. Chaos might be the reality.
You know how the Zs keep harping that socialism is the answer? Rather ironic that history has shown us exactly how socialism can “take care of” a 100-million citizen problem.
Chaos is one reality. So is slavery. Frogs patiently waiting for the boil, could easily be enslaved. Especially when they’re practically asking for it.
Re: (Score:3)
"Being employed is a human right"
It really isn't, and if you thought about it for a moment you would realize how stupid a thing to say that is. Being able to provide for yourself and your family is probably what you meant to say, but being employed? If only for the contradictions that being employed requires a employer, so what about the right of the employer to be employed?
Re: (Score:2)
The AI Convention mainly focuses on the protection of human rights of people affected by AI..
Great! So I take it you’re going to start by giving every human their lost-to-AI-speculation job back? No? Then stop bullshitting about your desire to protect human rights. Being employed is a human right, until governments stop colluding with Greed N. Corruption and define whatever UBI will be to replace human employment.
And it fucking better be an improvement on the current welfare system. Otherwise Greed won’t get a chance to remove that right from humans. Greed and Governments are going to have a major problem of their own making to contend with. If we can’t imagine how bad things would be with a 10% increase in unemployment, imagine the mass chaos of a much larger number.
It's not hard to imagine. Look at the way any economic downturn is handled in the United States. Poorer folks lose homes, businesses get bailed out, rich folks get tax breaks, and the poor get told they shouldn't have let it come to this and they can fuck off and fend for themselves. It's not going to change, and anybody counting on the government giving a shit as unemployment rises is completely disconnected from the reality of the world we live in. Now, I can't speak for other countries as I've only ever
The only important part is (Score:2)
Does it ban Terminators?
Re: The only important part is (Score:2)
It enforces Asimov's First and Second laws of robotics. Maybe.
Re: (Score:1)
No. In fact, it not only applies to every CPU with Branch Prediction given their crap definitions, but it explicitly excludes Skynet/WOPR(National Defense) programs and machines from the treaty.
Re: (Score:2)
"exemptions on AI systems used for national security purposes"
no
It legally binds YOU, but not themselves (Score:2)
treaty promotes responsible innovation and addresses the risks AI may pose.
exemptions on AI systems used for national security purposes, and limited scrutiny of private companies versus the public sector,
Isn't that clear enough that all those governments will continue irresponsible AI development regardless of the risk?
Who should anyone else bother with this so-called treaty?
Re: (Score:2)
Isn't that clear enough that all those governments will continue irresponsible AI development
AI isn't being developed by the government.
Re: It legally binds YOU, but not themselves (Score:2)
/me laughing in Pentagon and Chinese PLA.
Re: (Score:2)
Isn't that clear enough that all those governments will continue irresponsible AI development
AI isn't being developed by the government.
LOL. Yeah, you’re right. Why would they bother developing it when they can just take the technology when it’s ready. They can even take all the patents away and classify them under secrecy orders. Yes, even ones privately developed and owned There are thousands of US patents in this state already.
Ratification (Score:2)
Where is it and... (Score:1)
I don't trust them and I don't trust our government either. This probably has something to do with copyright and the jobs that AI is replacing.
Well, I don't agree with the logic behind Hollywood and I believe it's fair use if it's posted publicly or purchased one time by the company (no perpetual license crap). There is a reason the new AI models cost up to $2,000 / month according
Re: (Score:2)
So where is it and what does it say? I don't agree with secret "treaties" being ratified especially with the UK right now. I don't trust them and I don't trust our government either. This probably has something to do with copyright and the jobs that AI is replacing. Well, I don't agree with the logic behind Hollywood and I believe it's fair use if it's posted publicly or purchased one time by the company (no perpetual license crap). There is a reason the new AI models cost up to $2,000 / month according to OpenAI. Some of that is compute power, but some of that is because they have to pay for the data now, which is ridiculous. Now only millionaires and billionaires will have access to the technology, further making them even more wealthy and powerful and disproportionately hurting the middle class.
When you have millionaires and billionaires writing the laws and treaties and rules for the rest of us? What do you expect?
Then there is the jobs being replaced by AI and robots, also hurting the middle class. If it was a company, I'd say "adapt or die," but for people, that's a different story. UBI to some degree will have to exist if the majority of the middle class gets replaced with robots or AI. I mean, there are only so many people that would still have jobs. The profits of said companies will probably actually decrease because no one will be able to afford their products. This of course may usher in the mark of the beast and control of currency, as theft will run rampant in today's society. Especially if we deliver all checks on the 2nd Wednesday of each month like SSI. Direct deposit is better, but then you have loan companies gouging your paycheck with 600% interest. There's no winning for the needy.
I don't think anyone in the upper echelons of society is going to bother addressing the neediness of the people until it hits a point where it's affecting profits among the richer companies. By then, it'll likely be far too late to stem the tide of hate. Not sure what the outcome is when you have a very, VERY large mob of angry people with zero power, and a smaller, extremely powerful s
Oh? (Score:1)
It isn't legally binding unless it was ratified by a 2/3 vote of the United States Senate as required by Article II, Section 2 of the Constitution.
It isn't legally binding.. (Score:2)
But what if.. (Score:2)
..the AI doesn't agree to it?
Link (Score:3)
It is remarkably hard to find information on the actual contents of the AI Convention. This seems to be the web page describing it. [coe.int] Note the links to the actual documents.
The convention defines AI to mean "a machine-based system that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations or decisions that may influence physical or virtual environments." Wow, that's a broad definition!
The actual requirements of the convention reference human rights such as equality, non-discrimination and privacy "as enshrined in applicable international law and in its domestic law". All of the requirements are very vague, basically leaving the details to each individual country.
Another feel good, waste of time.... (Score:2)
nah it will have been watered down to favor (Score:2)