This past Friday, the New York Times published an op-ed by Oren Etzioni, head of the Allen Institute for Artificial Intelligence. The op-ed’s title is “How to Regulate Artificial Intelligence,” but the piece actually sheds little light on “how” legal systems could go about regulating artificial intelligence. Instead, it articulates a few specific “rules” without providing any suggestions as to how those rules could be implemented and enforced.
Specifically, Etzioni proposes three laws for AI regulation, inspired in number (if not in content) by Isaac Asimov’s famous Three Laws of Robotics. Here’s Etzioni’s trio:
- “[A]n A.I. system must be subject to the full gamut of laws that apply to its human operator.” For example, “[w]e don’t want autonomous vehicles that drive through red lights” or AI systems that “engage in cyberbullying, stock manipulation or terrorist threats.”
- “[A]n A.I. system must clearly disclose that it is not human.”
- “[A]n A.I. system cannot retain or disclose confidential information without explicit approval from the source of that information.”
These rules do a nice job of laying out the things that we don’t want A.I. systems to be doing. But that’s the easy part of deciding how to regulate A.I. The harder part is figuring out who or what should be held legally responsible when A.I. systems do those things that we don’t want them to be doing. Should we hold the designer(s) of the A.I. system accountable? Or the immediate operator? Or maybe the system itself? No one will argue with the point that an autonomous car shouldn’t run red lights. It’s less clear who should be held responsible when it does.
Etzioni’s op-ed takes no discernible position on these issues. The first rule seems to imply that the A.I. system itself should be held responsible. But since A.I. systems are not legal persons, that’s a legal impossibility at present. And other portions of Etzioni’s op-ed seem to suggest either that the operator or the designer should be held responsible. The result is a piece that punts on the real challenge of figuring out “How to Regulate Artificial Intelligence.”
I don’t have any issue with where Etzioni wants us to go. I’m just not sure how he thinks we’re supposed to get there.
In 2014, Elon Musk’s warnings about the dangers and risks associated with AI helped spark the debate on what steps, if any, government and industry bodies should take to regulate the development of AI. Three years later, he’s still voicing his concerns, and this weekend he brought them up with some of the most influential politicians in America.
In a speech before the National Governors Association at their summer retreat in Rhode Island, Musk said that governments need to be proactive when it comes to managing the public risks of AI:
» Read more
Credit: Randy Glasbergen
Here’s a quick roundup of law- and policy-relevant AI stories from the past couple weeks.
A British privacy watchdog ruled that a group of London hospitals violated patient privacy laws in sharing information with Google DeepMind. Given the constant push for access to data that all the major tech companies are making (in no small part because access to more data is crucial in the age of learning AI systems), expect to see many more data privacy disputes like this in the future.
Canada’s CTV reports on the continued push by some AI experts for “explainable” and “transparent” AI systems, as well as the skeptical response of other AI experts about the feasibility of building AI systems that can “show their work” in a useful way. Peter Norvig points to a potentially interesting workaround:
» Read more
Two opinion pieces were published this weekend–the second written in response to the first–on the issue of whether and how the rise of AI, robotics, and automation will affect another notable trend in modern society: economic inequality. Both authors make some intriguing points. But unfortunately, both also seem to have an unwarranted level of certainty about how AI will affect our economy and society.
» Read more
This is the final installment of a three-part series examining whether legal personhood is already possible under US laws governing limited liability companies (LLCs), which Shawn Bayern suggests provide an active path to personhood for autonomous systems. The first two posts in this series examined the two legal sources (New York’s LLC law and the Revised Uniform LLC Act) that Bayern used to support his contention that it is possible to use LLC laws to create an autonomous AI system with, for all intents and purposes, legal personhood.
The specific mechanism that Bayern proposed is creating an LLC whose operating agreement that effectively places the LLC under the control of an AI system, and then have every member of the LLC withdraw, leaving the system effectively unsupervised. I concluded from my own review of New York’s law and the laws of six states that have adopted RULLCA in some form that they do not provide a vehicle for creating LLC’s of the type Bayern described. The purpose of this final post is to examine a few other states’ LLC laws to see if my conclusions for New York and the RULLCA states are generalizable to other state laws.
» Read more
Last month, Technology Review published a good article discussing the “dark secret at the heart of AI”–namely, that “[n]o one really knows how the most advanced algorithms do what they do.” The opacity of algorithmic systems is something that has long drawn attention and criticism. But it is a concern that has broadened and deepened in the past few years, during which breakthroughs in “deep learning” have led to a rapid increase in the sophistication of AI. These deep learning systems operate using deep neural networks that are designed to roughly simulate the way the human brain works–or, to be more precise, to simulate the way the human brain works as we currently understand it.
Such systems can effectively “program themselves” by creating much or most of the code through which they operate. The code generated by such systems can be very complex. It can be so complex, in fact, that even the people who built and initially programmed the system may not be able to fully explain why the systems do what they do:
» Read more
This will, as it turns out, be a three-part series examining whether legal personhood is already possible under US laws governing limited liability companies (LLCs), which Shawn Bayern suggests provide an active path to personhood for autonomous systems. Bayern relied primarily on two sources of law: New York’s LLC statute, and the Revised Uniform LLC Act (RULLCA). Last week’s post explained why New York’s statute does not appear to provide a plausible path to AI personhood. This week’s will take the same critical approach to RULLCA and, more importantly, the states that have adopted some variation of RULLCA.
» Read more
Forewarning, this will be far longer and far more of a technical legal post than usual. It is also part 1 of what will be a 3-part post. Part 2 is posted here, and Part 3 is posted here.
One particularly hot topic in the world of law and AI is that of “artificial personhood.” The usual framing of this issue is: “should we grant ‘legal personhood’ to A.I. systems and give them legal recognition in the same way that the law recognizes corporations and natural persons?” This is, to be sure, an excellent question, and artificial personhood is one of my favorite topics to discuss and write about.
But some authors in the past few years, most notably Shawn Bayern, have gone one step further, claiming that existing laws already permit the recognition of AI personhood for all intents and purposes. Bayern focuses his attention primarily on the prospect of a “Zero-Member” or “memberless” LLC. (“Members” of a LLC are roughly analogous to partners in a partnership).
» Read more
Tom Toles, The Buffalo News, 1997
Last week I got an email from Will, an 8th Grader from Big D (little A, double L, A, S). He is in a class where the students get to choose a topic to write about, and he chose AI because he had “always wondered about what makes a machine better than humans in an area.”
Will emailed me wanting to know if I could answer some questions he had about AI and its impact on our society. I happily agreed, and he responded by sending five excellent questions. After getting approval from Will and his teacher (thanks, Ms. Peterson!), I am posting Will’s questions and my responses below. (I also sent Will an email with much shorter responses so that he wouldn’t fall asleep halfway through my answers).
Here they are:
What are your thoughts on the rapidly increasing investment in AI of huge companies such as Google and Microsoft?
This is one of the hottest topics in the world of AI policy right now. In some ways, the investment in AI by these companies is a good thing. There are so many things we could do with better AI systems, from having more accurate weather forecasts to reducing traffic on highways to helping doctors come up with better diagnoses when someone is sick. Those things would bring great benefits to lots of people, and they could happen much more quickly if big companies focus their time and money on improving AI.
On the other hand, there are always dangers when big companies get too much power. The usual way that we deal with those dangers has been through government action. But modern AI technologies are very complicated—so complicated that sometimes even the people who design them may not totally understand why they do what they do! It is hard to come up with good rules for things that no one completely understands.
» Read more
Credit: Mike Keefe
The ABA’s Science & Technology Law section has an AI and Robotics committee that holds a monthly teleconference “meetup” where a guest speaker presents on an AI/Robotics-related legal issue. From here forward, I’ll be making a brief post on each monthly meetup.
For the April meetup, Michele Kyrouz gave a presentation on California’s updated autonomous vehicle (AV) regulations. I wrote a post last fall discussing the new rules governing AV advertising and marketing, and intended to do a longer post discussing the regulation changes as a whole. This month’s meetup gave me the kick in the pants I needed to actually do that.
» Read more