AI & Law: Legal Apologies

The role of apologies via AI and the legal repercussions therein

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:

  • AI systems are increasingly being programmed to emit apologies
  • Some like the apology and feel better, while others find it insulting and a hollow gesture
  • There is a substantial body of legal literature on the topic of law-and-apologies
  • In some respects, an apology is legally exculpatory and in other ways legally inculpatory
  • Firms making AI might be digging themselves a hole, plus AI liability could later arise


Are you the type of person that expects to get an apology when someone else has done something wrong or that you feel otherwise owes you an act of contrition?

If so, you’ll be glad to know that AI systems are increasingly emitting a sorrowful response when they have done something that you don’t like or that has potentially hurt your feelings. That being said, keep in mind that today’s AI is not sentient and nor even close to reaching such a vaunted pinnacle. The apologetic AI being fostered onto the world currently is AI that has been programmed to emit an apology wherein the software developers have predetermined that an apology seems warranted.

For example, you apply for a car loan from an AI-based online system and the underlying Machine Learning algorithms determine that you are unworthy of being granted a loan. Rather than merely informing you that you have been rejected, the AI then also produces a kind-of apology that says the decision is quite regretful and hopefully you are not overly dismayed at the result. In essence, the AI is sorry to have cast you asunder.

Would such an apology soothe your bruised self-esteem and perhaps materially soften the abject disappointment at not getting the loan?

Maybe, at least that’s what the AI developers assume will happen.

On the other hand, some believe that the apology is akin to adding salt into the wounds of your denunciation. It is one thing to have a human provide a heartfelt apology, and altogether insulting to have a machine appear to be sorrowful. This seems like a hollow gesture and aims at those that do not realize that today’s AI is dense as a brick and utterly unlike human mental acumen.

Some decry this AI is fooling the public by seeming to have intelligence that it just doesn’t possess. As people get used to seeing these apologies, they will further anthropomorphize the AI and fall down a slippery slope of being deceived into believing that the AI is pretty much a humanoid.

For details on this and other AI and law topics, see my book entitled “AI and Legal Reasoning Essentials” at this link here:

The Legal Apologies Conundrum

From a legal perspective, there is another insightful point to be made.

The old maxim is that liability means never being able to say that you are sorry.

For those of you not especially familiar with the law-and-apology field of study, there is a storied history of how apologies, encompassing those made by humans and ostensibly by non-living entities such as corporations, have been treated by the law. One of my favorite recent articles covering this topic was authored by John Kleefeld and entitled “Promoting and Protecting Apologetic Discourse through Law: A Global Survey and Critique of Apology Legislation and Case Law” (appeared in Volume 7, Number 3, 2017 issue of Onati Socio-Legal Series).

In the United States, a landmark state law covering apologies was enacted in 1986 when a Massachusetts senator led and got passed a bill that was prompted by the fact that his daughter was killed by a car driver and the driver had said he wanted to apologize but was fearful of the legal consequences by doing so.

Researcher Kleefeld describes the apologies conundrum this way: “This view arises from the general rule that a party’s out-of-court statement or conduct against interest, not otherwise protected by privilege (e.g., the privilege associated with settlement discussions), is admissible against that party at trial even though it would otherwise be excluded as hearsay evidence. The Massachusetts bill sought to provide a ‘safe harbor’ for such would-be apologizers by rendering their sympathetic words or ‘benevolent gestures’ inadmissible in a civil action, and spawned comparative initiatives in almost every US state.”

Of course, lawyers and judges have been arguing about the nature of what constitutes an apology, doing so since perhaps the invention of adjudication. There are disputes over the scope and boundaries of an apologetic statement or utterance. This can be significant since it then shapes whether the apology gets into court or does not get included.

Generally, the rule-of-thumb is that an apology is composed of remorse, responsibility, resolution, and reparation, the so-called four Rs of legally defined apologetic expressions. If you merely tell someone you are sorry, this presumably is insufficient as a legally recognizable apology since it lacks or renders ambiguous at least three of the Rs.

Some assert that legislating the nature of apologies is good because it enables those that wish to apologize for the legal latitude to do so, and might therefore lead to the resolving of civil disputes in a more efficacious way. As per Kleefeld: “British Columbia’s Apology Act, or its model-act equivalent in Canada, the Uniform Apology Act, aims to do this for civil disputes generally. Other statutory provisions, notably in the US, aim to do this in the health care context.”

A reader of my column, lawyer Joseph McMenamin and a professor in the Department of Legal Medicine at Virginia Commonwealth University, brought to my attention the Virginia law encompassing health care providers and apologies, providing a handy example of such legal language, as stated in Virginia’s Death By Wrongful Act “the portion of statements, writings, affirmations, benevolent conduct, or benevolent gestures expressing sympathy, commiseration, condolence, compassion, or a general sense of benevolence, together with apologies that are made by a health care provider or an agent of a health care provider to a relative of the patient, or a representative of the patient about the death of the patient as a result of the unanticipated outcome of health care, shall be inadmissible as evidence of an admission of liability or as evidence of an admission against interest.”

Well, one thing to realize is that not everyone sees apologies as being quite so beneficial.

Some argue an apology is bound to set ablaze a matter that might otherwise not have risen to the level of being played out in the courts. The person receiving the apology is apt to feel that they got a clear cut acknowledgment of wrongdoing that was done to them, and therefore be spurred to file a legal case accordingly.

Another twist is that some laws make exculpatory the part of the apology that seems heartfelt, and establishes that inculpatory is the part that appears to be an admission of guilt. This then leaves juries with the impression that the apologizer seemingly admitted they were at fault and furthermore refused to express sorrowfulness for doing so.


In any case, we are likely to soon see that the AI-powered apologies open up another can of worms in the law-and-apology field. If there is any shift toward AI gaining a smidgeon of legal personhood, where then does the apology issued by an AI semi-autonomous or fully autonomous system land?

Being watchful of apologizing might be a savvy move, and the cutting remarks by Oliver Wendell Holmes, Sr. come to mind as added basis to avert proffering an apology: “Apology is only egotism wrong side out.”

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.

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