New AI laws have hidden controversies that the public needs to be hearing about.
getty
In today’s column, I examine the latest ruckus regarding proposed new AI laws that potentially let AI makers off the hook when it comes to having legal responsibility and accountability for potential harm caused by their AI.
Here’s the deal. Lots of new AI laws are being passed. In some of those laws, there are provisions regarding the scope and depth of responsibility and accountability that AI makers have for potential harm caused by their AI. If the provisions are lax and lenient, there are worries that AI makers will escape liability and be incentivized to take shortcuts on AI safeguards. On the other hand, if the provisions are overly strict, AI makers might hesitate to make advances to AI and be fearful that adding innovations will open a Pandora’s box of legal and financial exposures.
What is the prudent and best way to structure the legal and societal obligations that AI makers ought to meet? A drafted AI law in Illinois has sparked renewed debate on this difficult and harrowing question.
Let’s talk about it.
This analysis of AI breakthroughs is part of my ongoing Forbes column coverage on the latest in AI, including identifying and explaining various impactful AI complexities (see the link here).
As a quick background, I’ve been extensively covering and analyzing a myriad of facets regarding the intersection of AI and the law for many years. You can find my writings not only in my Forbes column but also as posted in Bloomberg Law, ABA Law Journal, The National Jurist, The Global Legal Post, Lawyer Monthly, The Legal Technologist, MIT Computational Law Journal, and so on.
There are two major perspectives on the mixture of AI and law:
Thus, you can apply the law to AI, and conversely, you can apply AI to the law. For my big picture overview of both of these exciting and rapidly evolving realms, see my discussion at the link here and the link here.
When it comes to applying the law to AI, the aim is to establish suitable regulations and provide appropriate governance on how AI should be devised and implemented. There are longstanding concerns that AI makers aren’t giving due attention to the ethical ramifications of their wares. Ethical issues are construed as “soft laws” and aren’t as formidable as legally enacted laws, known as “hard laws”. To level the playing field and keep AI makers on the up-and-up, some believe that we need more AI laws.
On the other side of the coin is the application of AI to the law. This consists of using AI to aid legal activities. Lawyers tap into the latest AI to devise legal strategies, brainstorm to find creative legal arguments, draft court filings, and prepare for cases by having the AI pretend to be an able adversary. For my extensive coverage on AI for legal reasoning (AILR), see the link here.
I’ve been diligently scrutinizing AI laws that are being enacted in the U.S. and throughout the globe, including:
In terms of the AI laws in the United States, they have not yet stood the test of time, meaning that we won’t really know how well they stand up until there are court cases that test these new laws. It is too early to know whether the laws will survive legal battles waged by AI makers and other contenders. Just because AI laws are enacted does not mean they are proper. All sorts of improper provisions and constitutionally contentious stipulations are undoubtedly buried within these shiny new AI laws.
Congress has repeatedly waded into establishing an overarching federal law that would encompass AI. So far, no dice. The efforts have ultimately faded from view. Thus, at this time, there isn’t an overarching federal law devoted to these controversial AI matters. The big question will be to what degree a sweeping federal law would impact the numerous state-level AI laws. The odds are that many of the state-level laws would run afoul of a federal mandate, and a tsunami of legal cases would arise as a tussle between federal law and state law is undertaken. It surely will be a legal mess.
The latest controversy about the wording of AI laws centers on a bill being drafted in Illinois, known as SB3444. It contains a provision that covers the possibility of AI causing critical harm. If AI does cause critical harm, an AI maker or developer is presumably on the hook for the actions of their AI. It is possible to word such liability in a myriad of ways. The wording can range from being lenient, middling, to firmly demonstrative.
Here is an excerpted portion of SB3444:
How does that phrasing strike you?
Let’s unpack those sentences.
First, as a warm-up, the excerpt emphasizes the circumstances under which an AI maker is not to be held liable.
This already poses an unstated hidden question. For example, is it better to stipulate when an AI maker is actually to be held liable, rather than to aim at when they are not liable? Some would insist that it is more encompassing to focus on the facets of when liability exists, not when it doesn’t exist.
Think of it this way. I tell you that if you drive my car, you won’t be liable for a car crash when it is raining or snowing. This seems reasonable. Suppose that a rainy day comes along. You get into the car and, while joyriding, scrape the car against the median. You would seem to be off the hook for any responsibility since I had agreed that you would be free and clear if rain were underway.
It might have been wiser of me to identify the conditions under which you are liable, rather than focusing on when you are not liable. It’s a delicate tradeoff.
Second, and the mainstay of acrimonious debate, the phrasing includes this passage: “if the developer did not intentionally or recklessly cause the critical harm.” Parse that into two separate conditions, namely, one of intentionality and the other of being reckless.
Intentionality stipulates whether the AI maker intended to bring forth harm. In the legal field, the act of intention is a big deal and encompasses the mental state of an actor, referred to legally as mens rea. To garner liability, you would have to substantively show that the AI maker wanted the AI to cause harm, or they consciously disregarded a known, significant risk. Anything else, such as an AI maker having made a mistake, or emergent behavior of the AI, and other facets, would largely fall outside the legal responsibility of the AI maker (all else being equal).
Recklessness traditionally means consciously disregarding a known risk. But with AI, a key question emerges: What counts as “known” when the system’s behavior is partly emergent? If a developer could have known through better testing, is that negligence or is it recklessness? The answer matters enormously. A narrow interpretation favors the AI makers; a broader one increases accountability.
Those who want the AI makers to be held to the strictest feasible liability would insist that the phrasing associated with “intention or recklessness” is overly simplistic and only covers a tiny iota of the range of potential responsibilities. They would liken this to calling out the tip of the iceberg but ignoring everything else that exists below the waterline.
The wording of SB3444 seemingly allows AI makers to be off the hook for way too much. Indeed, an argument to be made is that most real-world harms of AI exist in the sphere of incomplete foresight and imperfect safeguards. The phrasing doesn’t appear to capture that scope adequately.
Perhaps some suitable rewording or augmentation of the legal language could be undertaken to widen the scope to include what an AI maker should have known and done, while still giving a defensible path if they acted responsibly.
For example, here’s a possibility associated with reasonable care and foreseeable risk that might be considered that I’ve quickly crafted:
And, here’s a possibility associated with the underlying duty of care that I noodled:
Something along those lines might provide a more balanced semblance of AI maker liability. That being said, the wording would indubitably draw backlash from AI makers since there are ways to interpret the language such that it encompasses what might be construed as being suppressive toward advancing AI at a rapid pace. That takes us back to the crucial consideration of where a balance exists to provide sufficient protection while not inhibiting innovation.
For the moment, the general viewpoint is that this SB3444 bill, in its current formulation, seems like a wink-wink that favors the AI makers. It at least does have some semblance of stipulated liability, which is better than having nothing at all. But the clever wording appears to let the horse out of the barn. Once the horse is out of the barn, it is extraordinarily hard to get it back into the barn, akin to squeezing a genie back into the bottle.
For anyone who doesn’t closely keep tabs on the burgeoning set of new AI laws, the assumption is that these emerging AI laws are mindfully crafted to fully protect the public and ensure our safety and security when it comes to the advent of leading-edge AI. The reality is that the devil is in the details. Legal language can mask the reality of what the law will practically accomplish. Likewise, the legal language can allow gaps through which AI makers can legally avert or deflect responsibility and accountability for their AI.
Thomas Jefferson made this acclaimed remark: “You do not examine legislation in the light of the benefits it will convey if properly administered, but in the light of the wrongs it would do and the harms it would cause if properly administered.” Please keep that at the top of your mind when viewing any proposed AI law and even the already enacted AI laws. You’ll undoubtedly see the real truths hidden within the legal mumbo-jumbo.
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。