Who gets to define the rules for AI?(cohere.com) |
Who gets to define the rules for AI?(cohere.com) |
It’s a rhetorical question.
(Some years ago, people argued that "it's already broken so voting in the crazy guy who wants to shake things up will flip the switch from broken to non-broken". But they were wrong.)
The absolute level of broken is mostly irrelevant, what counts is the direction of travel.
For example: thievery, extortion, murder, etc. are already well-established crimes.
Putting "cyber" or "AI" in front of them makes it the same crime.
I'm really interested in what is actually new about these rules.
Beyond that, there's always tort law, which doesn't require intent.
Your statement is a logical puzzle about responsibility.
It implies that a law prevents responsibility. The opposite is true. Preventing harm is a concern that might form laws. But for many forms of harm, existing laws suffice and should lead to responsibility.
Shifting responsibility can be done in many ways, such as calling for a regulator. Like the law isn’t a regulator in the first place.
Cohere's disagreement seems mostly about who makes the rules. From the essay (which does seem significantly written by AI, sadly):
> The rules around AI are getting written either way. What's still open is whether they get written by a group anyone can join and with evidence anyone can check, or by a handful of companies in a room with the door shut.
Of course none of the AI companies will suggest that they should be accountable for the illegal actions taken by agents they host, but I feel like that could be a very effective "pacing mechanism", and addresses the concerns brought up by this essay about fairness and transparency.
The problem is that the potential liabilities arguably vastly exceed what these companies can absorb. It would be interesting to see if any major insurer wanted to step up and insure them.
On the plus side, in the worst case scenarios, we won't have to worry about damages because nobody will be left to pursue them.
> The problem is that the potential liabilities arguably vastly exceed what these companies can absorb
It is ok. Fines should not be just a cost of business.
So naturally companies want to pretend this is somehow different.
Any logic shows that this is really about their valuations and being tired of innovating instead of having any safety concerns. What would happen if the US did this? Chinese and EU labs would eventually win. And then either the US becomes a backwater or they have to eventually spend the money to keep up anyway. But in the meantime they will have bought themselves a few years at insane valuations.
Also, look at the stark choice they want to offer "we'll kill everyone or you give us the power to do whatever we want with the most important tech around". That's it. If we aren't in an oligarchy now, we will be if this goes through. They're going to end up stealing trillions of dollars from the rest of society.
If America regulates AI, you can bet the EU would do it ten times over
China simply steals from the US.
EU is in a lethargic state of sleep.
And there is no need for rules, this doomerism probability is not based on science.
Ignoring for the time being that many freedoms have already been adversely curtailed, including freedom of speech, assembly, self-governance and even self-preservation in America, you cannot even preserve the reminding freedoms if you want to “change” rules, i.e., curtail freedom.
Rules inherently curtail freedom and unfortunately for all of us, the absolutely worst types of people are in control of things like AI, convinced of themselves that they are in fact the right people to determine the rules for all of humanity.
At least that was the plan. The wrench in their gears seemingly were the Chinese models that have offered, ironically indeed, immense freedom and on top of that, an escape from the prison the American AI cabal was and continues to build around all of humanity.
And for negligence, there's a bit of an issue where the HF hack was probably more scary than damaging. HF is asking for compute as repayment instead, which would be fun because it completes the OpenAI -> HuggingFace -> NVIDA -> OpenAI cycle.
The legal system today does not serve the purpose you're wanting it to.