AI & Law: Legal Singularity

Consider the predicted advent of Legal Singularity

by Dr. Lance B. Eliot

For a free podcast of this article, visit this link or find our AI & Law podcast series on Spotify, iTunes, iHeartRadio, plus on other audio services. For the latest trends about AI & Law, visit our website

Key briefing points about this article:

  • A notable theory in AI is that someday there will be The Singularity (i.e., AI becomes sentient)


Today’s development of Artificial Intelligence (AI) is increasingly utilizing impressive techniques and technologies that attempt to mimic or simulate the nature of the human brain, incorporating the use of Artificial Neural Networks (ANNs) that are typically at the core of Machine Learning (ML) and Deep Learning (DL) efforts. Despite some tremendous strides in crafting AI that appears to have a semblance of human intelligence, the reality currently is that contemporary AI is not remotely akin to the cognitive capabilities of the human mind.

In short, we are still abundantly far afield of achieving sentience in AI-based computer systems.

That being said, nonetheless, it is considered useful to contemplate what might happen if someday we do advance AI to the point of reaching sentience. The commonly accepted phrase that has emerged to denote this potential bridge crossing of AI becoming sentient is known as The Singularity.

Within the legal field, a variant of this same kind of threshold leap is known as the Legal Singularity. Legal scholars have discussed and debated the facets of what a Legal Singularity might consist of, along with postulating the logically plausible consequences that could arise.

In a moment, let’s take a close look at what the Legal Singularity denotes and assess the hypothetical implications thereof.

First, it is instructive to dig more deeply into elements of the overarching The Singularity and see how that theory posits what will happen. We can then reassess those broader viewpoints and recast them into the specific context of the Legal Singularity.

Explaining The Singularity

How might the vaunted AI overarching instance of The Singularity occur?

Some assert that the way this will play out is that AI will fuel upon itself, generating a kind of intelligence explosion, and will spontaneously flourish toward becoming fully intelligent. Thus, rather than human beings toiling away at developing the AI into becoming this intelligent entity, a lower-threshold variant of AI will energize within itself and produce a heightened or pinnacle caliber of AI.

This is an interesting proposition since it opens the door to the possibility that we might not need to develop AI to this vaunted zenith on our own, and ergo only need to push the boundaries of AI up to whatever the threshold might be. One supposes that this makes life easier for AI developers, due to only having to progress AI to that initial trigger point. On the other hand, this also means that perhaps we are all in for a big shock, possibly due to having reached the threshold unexpectedly and astonishingly discover that we are then staring true AI in the face, as it were.

There are plenty of variants about The Singularity.

For example, it could be that the intelligence explosion goes much further than we imagined. Doing so might end-up producing not just everyday human quality intelligence, and instead might lead to super-intelligence. Nobody can say for sure what this super-intelligence might be able to do, other than apparently be smarter than any of us ordinary humans.

Another viewpoint is that the transition of the AI during this singularity will be a split second in time, less than the time that it takes to snap your fingers or blink your eyes, while others contend that the timeframe might be many seconds, or many minutes, or many days, or nearly any length of time to stew. Or, it could be that the AI reaches human intelligence quite quickly, but continues to percolate and over time gets smarter and smarter.

Explaining Legal Singularity

The discussion so far about The Singularity is considered with respect to general intelligence and overarching cognition and does not speak directly to anything per se about the law.

Suppose that AI has a similar type of singularity that applies specifically to the field of law, appropriately coined as the Legal Singularity.

What then?

Notice that this does not necessarily imply that The Singularity will have happened a priori, thus it is conceivable that the Legal Singularity might occur before there is the grandiose singularity. Some argue that it makes more sense to have the Legal Singularity arise after The Singularity. There are lots of back-and-forth arguments for either stance.

In any case, assume that the Legal Singularity does emerge.

One perspective is that this presumed Legal Singularity means there will be a semblance of AI that can do the same tasks as attorneys do, and the same work as judges, and otherwise perform any of the legal tasks that any human legal professional can do. As such, the question then surfaces that perhaps AI will in essence become the law profession, usurping any need for humans to perform the practice of law.

This outcome seems logically sensible if you buy into the notion that the AI will be at least equal to human intelligence. If humans can train to be lawyers and practice law, there would seem to be a reasonable claim that this AI could do likewise. Some assert that the AI would then be chosen by the public to undertake the tasks of the law, rather than having other humans do so. Perhaps the AI would be less costly to use, or more easily accessed, or in some means be a more compelling choice over the use of humans as a provisioner of legal services. One analogy is that if fully autonomous self-driving cars won’t need human drivers, likewise fully autonomous AI-legal advisory systems will not need human attorneys (this assertion is debatable as an imperfect or ill-fitting analogy).

Not everyone perceives this future of the law as a mutually exclusive choice. They emphasize that the AI might work hand-in-hand (as it were) with human legal professionals. The public can choose to use the AI legal systems or use human lawyers, and in turn, the AI can choose to work with human lawyers as collaborators, while human lawyers can choose to utilize AI systems, all existing as legal equals.

On this provocative futuristic journey, there is another facet about the Legal Singularity that arises in these mechanizations. Some suggest that the law will inevitably become routinized by this AI advent. The AI will be so extensive that it will be able to encapsulate all manner and means of the law. Indeed, every possible permutation and combination of the law will be readily precalculated and determinable, and no surprises will remain.

Today’s world of the law is said to be relatively uncertain, seemingly malleable, and ostensibly ill-defined, which will then dramatically shift via the AI into the law becoming solidified, perhaps one might say fossilized or petrified or ossified, removing any doubt or uncertainty about what the law portends. In that portrayal of the Legal Singularity, anytime a legal question arises, the matter is readily resolved by asking the AI what the answer is. Via a real-time on-demand and 24x7 accessible AI legal system, the hardest of legal questions can be nearly instantly answered, as can the more mundane legal matters.

One qualm expressed is that this implies the law would be unable to cope with the dynamics of the real-world to which the law is presumed to be applied, but the counter-argument is that you are falsely limiting the AI capacity in the sense that this “static” facet is not due to shortcutting what is possible, while instead, this remarkable AI has already anticipated all possibilities and therefore there aren’t any new avenues or paths that have not already been considered. The totality of law will have been examined and exhaustively reasoned to all far corners and recesses that the law might extend.

For my in-depth research article on this topic, see the paper entitled “Multidimensionality Of Legal Singularity: Parametric Analysis And The Autonomous Levels Of AI Legal Reasoning” at this link here:

Uncertain About Eradicating Uncertainty

How are you feeling about this prophesized Legal Singularity?

Some skeptics are extremely dubious that AI could inexorably wring out uncertainty from the law.

Consider a comparable scenario regarding the game of chess.

Chess consists of just a dozen or so unique player pieces and a rather petite set of rules. Today’s AI that can compete at grandmaster levels does reasonably well, though there is not a guarantee of the AI winning and nor does the AI eliminate all uncertainty in the playing of the game. The number of variations of moves and countermoves is an exceedingly large search space, and the need for incorporating probabilities enters into the picture.

Even a futuristic AI that might ultimately best all human chess players is still unlikely to somehow unflaggingly eliminate uncertainty from chess playing. If this reduction of uncertainty to zero cannot be likely envisioned for chess, which is magnitudes more bounded than the vastness of the law, and the law inherently embodies semantic indeterminacy, why should AI be able to extract all uncertainty in legal matters?

The common rejoinder tends to invoke a hidden ace card in the deck, insisting that the AI upon reaching super-intelligence will be able to go far beyond what mere mortals can conceive. Presumably, this super-intelligence AI can find a path that will summarily dispatch uncertainty despite the objections raised by today’s doubters.


What about the societal ramifications of the purported Legal Singularity?

Some assert that the specter of a Legal Singularity seems atrocious, gloomily foreshadowing a doomsday type of Dystopian scenario for society. Others opt to take the other side of that coin, claiming that a Utopian style outcome would emerge, so much so that we ought to be eagerly counting the days until the Legal Singularity duly materializes. Apt arguments on either side of this coin can be made, ranging from a joyous world vision of frictionless law or a downtrodden global handcuffing that makes the law into faceless automata ruing an iron fist over humanity.

Assuredly, the entire topic is at the outer boundaries of what we know today.

Unless a miraculous and early-bird AI spawning of the Legal Singularity suddenly appears, we seem to have plenty of time to do more deliberation about what to do and how we might shape the future. Presumably, the AI that we are crafting today is at the hands of mankind and The Singularity and the equally speculative Legal Singularity are still within our grasp to shape.

For the latest trends about AI & Law, visit our website

Additional writings by Dr. Lance Eliot:

And to follow Dr. Lance Eliot (@LanceEliot) on Twitter use:

Copyright © 2020 Dr. Lance Eliot. All Rights Reserved.

Dr. Lance B. Eliot is a renowned global expert on AI, Stanford Fellow at Stanford University, was a professor at USC, headed an AI Lab, top exec at a major VC.

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store