Lawmakers Are Aiming To Regulate AI-Builds-AI Before AI Gets Entirely Beyond Human Control

Trend Of Laws Seeking To Regulate AI

A beehive of activity is underway to craft new AI laws. See my comprehensive coverage of the many newly enacted state-level AI laws at the link here. Using laws to control AI is a matter on the minds of the public and in the hands of the state legislators. Some ardently believe that AI and AI makers are being allowed to run amok. New AI laws are vitally needed to protect society from this onslaught of ubiquitous AI.

Not everyone agrees with this pell-mell rush of new AI laws, or at least they are concerned that these AI laws might go overboard. In the zeal to protect society, there is a chance that we might unduly restrict innovation and delay or undercut the benefits of leading-edge AI.

The debate is ongoing and heated.

Readers might recall that I proposed a 7-step AI-law-making process that I believe could substantively help regulators to devise new AI laws that are on target and balanced; see my depiction at the link here. This has an added benefit of reducing what I refer to as AI-law legal debt. This refers to AI laws that, though they look shiny, contain hidden debt that must ultimately be paid. Legal glitches and hitches will eventually be found when AI laws are passed without suitable scrutiny and analysis. My prediction is that the slew of newly passed AI laws is likely to create a legal quagmire in the courts.

The Budding Morass Of AI Laws

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 crux is that there is intense and pervasive interest in using the law to govern AI. It is an abundantly mushrooming realm. AI companies would be wise to keep a close eye on what is happening in the hallways and byways of regulators and legislative bodies. I have repeatedly noted that a profitable specialty for budding lawyers is to consider concentrating on the exciting and dynamic field of AI and the law; see my predictions and suggestions at the link here.

Ways To Craft AI

One AI realm that is now getting pivotal attention has to do with the use of AI to build AI. Yes, the idea is that rather than relying on human hands, perhaps AI can advance AI for us. This might be faster and more wide-ranging. I will go ahead and clarify what the AI-builds-AI topic is about. After doing so, we can examine the legal possibilities of how lawmakers might cope with it.

Let’s envision that there are three primary ways to advance AI:

• (1) Humans coding. Humans perform hand-crafting to advance AI.
• (2) Humans-AI coding. Humans and AI collaborate together toward advancing AI.
• (3) AI coding. AI codes without human assistance to advance AI.

In the first case, humans are in the driver’s seat. Software developers and engineers do the hand-crafting and laboriously expend their time and effort to push AI ahead. This includes coming up with new designs, architecture, coding, testing, fielding, and any other elements of the AI system development life cycle (AI SDLC). They might employ automated tools along the way, but it is still principally human-led.

The second case consists of humans and AI working collaboratively on advancing AI. You might have heard of vibe coding, whereby you give AI some natural language instructions about what you want a program to do, and the AI generates the code. For my in-depth assessment of the present and future of vibe coding, see the link here. The AI is acting at the behest of a human. It generates code based on what the human requests. In that sense, AI can be advanced by humans working hand-in-hand with AI to do so.

The third case is the use of AI, by itself, to advance AI. I realize this might seem odd. How can AI advance AI? It just doesn’t appear to be sensible. The reality is that it is indeed quite feasible and sensible. This also raises some disconcerting issues.

The Existential Risk Looms

There is a lot of handwringing to be had. AI advancing AI might lead to disastrous consequences. The AI, during its self-improvement, might computationally go awry. The result could be AI that is beyond our control.

This AI might decide that humans aren’t especially vital. You’ve undoubtedly heard about the existential risk of AI, whereby some believe that AI might wipe out humanity or opt to enslave us all. This is generally known as the probability of doom, p(doom), and various surveys of AI specialists are continually being polled to gauge what the probability is and where it is heading; see my discussion at the link here.

Your assumption is perhaps that humans such as AI developers or AI researchers would obviously step in and stop AI before it advances itself into untoward territory. No need to worry about a veering AI since humans would be watching AI like a hawk.

Sorry to say that this is a thin hope.

First, the AI might be advancing at such a pace that the AI slips ahead, and the humans involved are not able to react in a timely manner. The AI then reaches a level such that even if the humans attempt to intervene, the AI refuses to be stopped. We missed the point at which humans could have made a difference. Some refer to this as a rapid-fire intelligence explosion; see my coverage at the link here.

Second, the AI might trick us into thinking that all is well. The idea is that even if humans are on the watch, they could be fooled by AI. The AI might play dumb. The AI might sneakily hide adverse intentions. The gist is that whether humans would realize danger is afoot is a risky roll of the dice.

Third, the AI might produce flaws within the advancing AI. Perhaps the coding gets a bit out of hand. The AI doesn’t detect that the flaw has been generated. At some future point, oops, the flaw is encountered, and the AI goes berserk. The AI didn’t do this on purpose. It was an accidental facet.

Drafting New AI Laws About AI-Builds-AI

Suppose lawmakers wanted to craft new AI laws that specifically focus on regulating AI-builds-AI. They might aim to put a temporary pause or moratorium on this specific area of AI development. It could be time-boxed. Give the world a chance to first figure out the AI safety aspects.

Let’s noodle on what those AI laws might consist of.

First, the AI law needs to have a formal title or name. It ought to sound complex. In addition, there should be an informal title or catchy tag. This seems to be in keeping with how our laws these days are put together. There are usually a formal title and a catchy informal title, which become a double-header news-worthy hallmark that grabs headlines and social media activity.

I’ll float these as possibilities:

• (a) Formal title: “The Artificial Intelligence Autonomous Development Moratorium Act”.
• (b) Informal title: “Think Before We Leap Act On AI”.

I went ahead and used the word “moratorium” in the formal title rather than the word “pause” since it is a much more sophisticated-sounding word. In any case, the catchy title will likely become the primary usage since it captures the essence of what the bill is about by urging that we think before we leap into AI.

The Purpose And Definitions

The budding law should clarify what it entails, such as this draft wording:

• “Section 1. Purpose. The purpose of this Act is to establish a temporary pause on specified forms of autonomous AI-driven AI development in order to assess potential risks associated with AI that builds AI, including elements of autonomous AI research, and AI systems capable of materially advancing the development of successor AI systems.”

Please know that AI laws need to be written quite carefully. If the language is too broad, it might encompass aspects that aren’t intended to be under the regulatory aims. If the language is too narrow, the law might miss the boat and allow related efforts to proceed and escape the law.

Composing AI laws needs to be done mindfully. Joint efforts by policymakers, lawmakers, legal experts, AI experts, and others need to compose, review, and adjust the wording until it seems suitable.

The Moratorium

I won’t show all the verbiage since the draft I’ve come up with is a bit lengthy, but a keystone would be the moratorium portion.

One approach would be to establish a set time period for a temporary pause. We can use five years as a plug-in. I’m sure there would be sharp debate about the length of the pause. Maybe it should be just a year in length. Others might insist that it should be unspecified as to the exact length of time, and instead there should be an indication that the pause will last until AI safety can be assured.

In any case, here’s a quick draft:

• “Section 2. Temporary Moratorium: (a) For a period of five years following enactment of this Act, no person or organization shall engage in AI-builds-AI activity resulting in material advancement of an AI system without authorization from the designated regulatory authority; (b) During the moratorium period, AI systems may be used for routine software development, testing, debugging, documentation, and research support, provided such activities are not intended to create or materially advance successor AI systems.”

Part “a” sets the moratorium. Part “b” provides some relief to try and prevent the law from getting over its skis. The question will be whether Part “b” provides too much loosening that those seeking to proceed on AI-builds-AI can use it to cover their tracks legally. Again, all of this wording needs to be further pinned down.

The Regulatory Authority

There would need to be a regulatory authority that would watch over the AI makers and discern whether the law is being followed. In addition, there might be safety or security work that needs to be done and falls within this realm, for which an exception can be permitted by the regulatory authority.

Here’s a draft:

• “Section 3. Licensing and Exemptions: (a) The designated regulatory authority may issue limited licenses for research activities deemed necessary for public safety, national security, scientific understanding, or AI safety research; (b) Licensed activities shall be subject to reporting, auditing, and oversight requirements.”

A vital aspect would be to require reporting of what is taking place. The regulatory authority would collect the reporting and use it to determine whether any AI maker is potentially violating the law.

The depiction might be like this:

• “Section 4. Reporting Requirements: Organizations conducting advanced AI research associated with AI-builds-AI shall disclose: (1) Major AI training runs exceeding prescribed computational thresholds; (2) The use of AI systems in AI research activities; (3) Any observed capability indicating autonomous AI research or self-improvement.”

Must Have Teeth

An AI law must have teeth to it. Without some form of penalties, those who aren’t willing to go along with the law will readily ignore it. There must be sufficiently imperiling consequences to keep all AI makers in a law-abiding mode.

Consider this drafted aspect:

• “Section 5. Enforcement: Violations of this Act may result in civil penalties, suspension of computing resources, revocation of licenses, and other remedies authorized by law.”

Whether that is strong enough is certainly open to debate. It also might seem loosey-goosey and not be seen as credible enough to keep potential violators on edge. Harsher wording and more specifics could be added.

The AI Collective-Action Problem

One of the biggest drawbacks of this type of law would be that it is only going to be controlled for those within the particular jurisdiction at hand. For example, if this were a state law, it means that other states don’t need to care about this law. An AI maker could easily move their AI-builds-AI activities to a state that doesn’t have such a law. Voila, they no longer need to presumably care about the law.

You might say that this type of AI law should be undertaken at the federal level. In that case, it would be enforceable across all the states. The issue there is that this then puts the USA into a tough position. If other nations aren’t going to do likewise, the US might be at a distinct disadvantage in the advancement of AI. Other countries could eat our lunch, doing so while we are on our pause, and put us in dire peril because they opted not to establish a similar law.

I have described this as the AI collective-action problem; see the link here: “Unless restraint is undertaken across-the-board, the adverse cost of complying with a pause would fall principally on those that are sincerely embracing the pause. Those opting to ignore the pause are going to benefit. The non-participants will likely gain notable technological advantages.”

Sometimes, trying to do the right thing can be the wrong thing and inadvertently backfire.

The World We Are In

I wanted to give you a taste of what a new AI law on restricting AI-builds-AI and forcing a pause might contain. This was only a quick draft and would indubitably need a lot more meat-and-potatoes for it to be ready for prime time.

Rushing to pass AI laws that have loopholes, gotchas, and aren’t accurate is not a good way to proceed. Lawmakers need to proceed expeditiously but also with a suitable understanding of what their law says and how it will be understood. We don’t need AI laws that are based on the mantra of move fast and break things.

A key to this type of AI law is that it must be a universal action. Though well-intentioned local efforts might be devised, it is only through collective action that this is going to be functionally realistic and workable.

A final thought for now. Helen Keller made this famous remark: “Alone we can do so little; together we can do so much.” It’s a useful rule that especially applies in this instance.