AI & Law: Soft Law About AI

Original Source Here

AI & Law: Soft Law About AI

A vital collection of soft law about AI has been posted by ASU

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:

  • Soft law is oftentimes a precursor to officially formalized hard laws
  • The arising of soft law tends to occur when shoring up new domains or innovations
  • A slew of soft law has emerged in the burgeoning field of AI
  • Arizona State University (ASU) College of Law has established a database of soft law about AI
  • This database is a veritable treasure trove, publicly available, and can be sliced-and-diced


Some would contend that there is the law and then there is everything else. You’ve likely heard that well-worn line before. It is certainly eye-catching and memorable.

What makes the catchphrase especially interesting is that somewhere in that morass is so-called soft law, sitting somewhat in a no man’s zone. Yes, there is a nebulous grey area that is not quite a law and yet oftentimes provides a law-like shaping and tonal directive toward what we can do, including whether our actions are seemingly lawful or ostensibly could veer into becoming unlawful. This mushy realm is commonly known as soft law.

It is said that soft law at times provides the seeds for the formation of future on-the-books hard laws. In a sense, soft law is floated around and if it gets enough traction then the odds are that official regulations and more formal law will be enacted based on those tried out precepts.

The malleability of soft law is a great strength and simultaneously a potential weakness. As a strength, soft laws can be changed and refined readily. No big hassle, no need to jump through a lot of arduous hoops. Adjustments can be made, remade, and altered time and again. You could argue that this is a big plus since soft law can iteratively be improved until it is finally ripened to be turned into conventional law.

A downside is that all that malleability means that there is something amorphous and intangible about soft law. You cannot stake all your eggs on the soft law since at any time it can be revised, revamped, or entirely reimagined. That is a bit of a slippery slope to depend upon.

One of the most frequent growth patterns of soft law entails domains that are new to the conventional forces of law. If there are scant on-the-books laws in the legal basket about a new domain then the advent of soft law can rush in to momentarily bolster the fresh ground. Rather than waiting for the mainstay of law to catch up with a rapidly advancing innovation, soft law speedily emerges as a stopgap to signal what might be coming down the pike later on.

Soft Law About AI

What innovation or domain comes to mind as a key exemplar of this soft law burgeoning appearance?

A somewhat new innovation that has garnered a tsunami of soft law is the modern-day emergence of Artificial Intelligence (AI) as evidenced by the myriad of AI-based apps and AI-infused computer systems that keep being unleashed. They are seemingly everywhere. And you can bet your bottom dollar that the wild crush of more AI is going to steam ahead, relatively unabated.

There are two ways to think about the law and AI.

I’ll briefly highlight the two approaches herein. For my detailed analyses on this and other AI and law topics, see my book entitled “AI and Legal Reasoning Essentials” at this link here:

One approach is to consider how the law ought to govern or regulate the development, testing, and fielding of AI systems. Some fervently contend that existing laws are insufficient to handle what today’s AI is bringing to the world, including qualms about embedded biases and other maladies.

The second approach involves applying AI to the law, potentially automating aspects of the law such as being able to computationally undertake legal reasoning. An overarching goal for many entails someday producing autonomous AI systems that can perform legal services on par with human lawyers. That’s the famous or perhaps infamous lawyer-in-a-box efforts that are energetically arising in the LegalTech marketplace.

We will focus this specific discussion on the first approach, dealing with laws that govern the advent of AI. In a manner of consideration, you could point out that the governance of the AI system is also applicable to the second approach, wherein we ostensibly need laws to aid in governing the application of AI to the law and aim to ensure that AI Legal Reasoning (AILR) is purposefully overseen.

As mentioned, many emphasize that existing laws are inadequate when it comes to overseeing and regulating the AI tidal wave. Readers of my columns might recall that I’ve covered this topic in various ways.

A newly released collection covering soft law about AI has become available and I would earnestly applaud this as a sorely needed and altogether handy treasure trove, especially since it is publicly accessible for all to see and leverage. Esteemed colleagues Carlos Ignacio Gutierrez and Gary Marchant at Arizona State University (ASU) College of Law have made a herculean effort to put together this amazing database (which can be accessed at along with research papers describing the collection, and heartily acknowledges the many ASU law students that aided in the compilation and screening process).

Their jointly co-authored paper states this about the work: “Soft law is defined as a program that sets substantive expectations, but is not directly enforceable by government. Because soft law is not bound by a geographic jurisdiction and can be developed, amended, and adopted by any entity, it will be the dominant form of artificial intelligence (AI) governance for the foreseeable future. The objective of this document is to compile and analyze global trends on how this governance tool is used by government, non-profits, and the private sector to manage AI’s methods and applications.”

A resultant database in a spreadsheet is posted online and contains information about a wide array of soft law AI initiatives. You can easily slice-and-dice the database via the spreadsheet, such as grouping by country of origin, type of entity, and the like. There are over 600 soft law AI efforts listed. The researchers also sought to categorize the set by various themes and sub-themes: “One of the original contributions of this research is its classification of each program’s text. By harnessing our 15 themes and 78 sub-themes, we uncovered that general mentions of transparency, general mentions of discrimination and bias, and AI literacy are the most represented issues in our database.”

Anyone interested in soft law about AI, and that is trying to garner too what formal regulations might eventually look like, should peruse this collection. Per the existing need to try and rein in the AI out-of-control spiral of AI For Bad (meaning adverse outcomes as sparked by AI systems), soft law is trying desperately to curtail the numerous bad apples in the barrel.


As a final remark, for now, the researchers make this poignant indication: “Soft law is not a panacea or silver bullet. By itself, it is unable to solve all of the governance issues experienced by society due to AI. Nevertheless, whether by choice or necessity, soft law is and will continue to play a central role in the governance of AI for some time.”

That is a keen insight and showcases that sometimes soft law is the only sheriff in town, while the rest of the troops are gradually, slowly, painstakingly getting formed up.

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 © 2021 Dr. Lance Eliot. All Rights Reserved.


Trending AI/ML Article Identified & Digested via Granola by Ramsey Elbasheer; a Machine-Driven RSS Bot

%d bloggers like this: