WSJ logo
Forbes logo
Fox News logo
CNN logo
Bloomberg logo
Los Angeles Times logo
Washington Post logo
The Epoch Times logo
Telemundo logo
New York Times
NY Post logo
NBC logo
Daily Beast logo
USA Today logo
Miami Herald logo
CNBC logo
Dallas News logo

Professional Liability and AI: When Can (and Should) Professionals Rely on Artificial Intelligence?

While artificial intelligence (AI) presents unique challenges for individuals and organizations in many industries, this is perhaps nowhere more true than in the professional service industry. Currently, professionals must practice risk management and be careful to avoid using AI to substitute for their professional judgment—especially when AI gets it wrong. But there are also legitimate questions as to when it may become malpractice for professionals not to rely on AI.

So, as a licensed professional, when can (and should) you rely on artificial intelligence? As we have already intimated above, this is not an easy question to answer. Relying on AI may be appropriate in some circumstances but not others (and may be appropriate for some professionals but not others), and whether reliance on AI gives rise to professional liability will generally be determined on a case-by-case basis.

Using AI in Professional Practice: The Importance of Verification

With the current state of AI, in virtually all circumstances, professionals who choose to use AI in their practices should still independently verify the accuracy of any research, written material, or other outputs generated using AI platforms. In recent years, we have seen several fairly embarrassing cases of professionals and other entities relying on AI to conduct research, generate work product, and provide advice when they shouldn’t have. For example:

  • A judge fined two lawyers and their law firm for submitting briefs that contained fake case citations generated by AI.
  • A venture capital firm lost a deal after an AI meeting assistant created and automatically disseminated a transcript of a meeting that included its investors’ post-meeting discussions of the firm’s “strategic failures and cooked metrics.”
  • New York City’s MyCity chatbot gave entrepreneurs inaccurate legal advice, including “claim[ing] that business owners could take a cut of their workers’ tips, fire workers who complain of sexual harassment, and serve food that had been nibbled by rodents.”
  • An AI-based tool for diagnosing patients’ medical conditions was found to be seriously flawed due to the nature in which it was trained. Among other issues, the tool “learned to identify COVID risk based on the position of [a patient] in [a] scan,” based on the fact that patients who are lying down tend to have more serious medical conditions than those who are standing up.
  • In another healthcare-related case, an AI algorithm used to predict patients’ medical treatment needs was found to be biased against Black patients, causing Black patients to receive substandard care.

As the judge wrote in the case involving fake AI-generated case citations, “Technological advances are commonplace and there is nothing inherently improper about using a reliable artificial intelligence tool for assistance. . . . But existing rules impose a gatekeeping role on attorneys to ensure the accuracy of their filings.” This is true not only in the legal profession (and not only with regard to filings), but in other professions (and in relation to other professional activities) as well. The mere existence of AI does not alter professionals’ duties of competence or other ethical obligations in any industry. So, while professionals can generally use AI tools to assist with their professional practice, they cannot use AI tools to perform their professional practice for them.

At a bare minimum, this requires verification. In each of the cases above, the AI-related risks involved could have been avoided had the professionals who were involved (or who should have been involved) examined the relevant AI outputs. Lawyers can easily verify whether case citations are real, and lawyers also could have easily spotted the flaws with the MyCity chatbot. Simply reviewing the meeting transcript in the venture capital firm case would have saved the firm’s deal. Regarding patient safety, medical diagnoses, and health care delivery, while AI can help point doctors in the right direction, doctors should not currently be basing diagnostic decisions solely on what AI tells them.

Put our highly experienced team on your side

Dr. Nick Oberheiden
Dr. Nick Oberheiden

Founder

Attorney-at-Law

Lynette S. Byrd
Lynette S. Byrd

Former DOJ Trial Attorney

Partner

Brian J. Kuester
Brian J. Kuester

Former U.S. Attorney

Kevin McCarthy
Hon. Kevin McCarthy

55th Speaker, U.S. House of Representatives (ret.)

Government Consultant

Mike Pompeo
Mike Pompeo

Of Counsel

Former U.S. Secretary of State

John W. Sellers
John W. Sellers

Former Senior DOJ Trial Attorney

Linda Julin McNamara
Linda Julin McNamara

Federal Appeals Attorney

Nicholas B. Johnson
Nicholas B. Johnson

Former Prosecutor

Roger Bach
Roger Bach

Former Special Agent (DOJ)

Chris Quick
Chris J. Quick

Former Special Agent (FBI & IRS-CI)

Michael S. Koslow
Michael S. Koslow

Former Supervisory Special Agent (DOD-OIG)

Ray Yuen
Ray Yuen

Former Supervisory Special Agent (FBI)

Professional Liability Concerns Related to AI Use

The importance of verifying AI outputs and maintaining independent professional judgment is apparent when we look at the variety of issues that can lead to AI risks—and potentially justify professional liability or malpractice claims. These issues include (but are not limited to):

Inaccuracies and Bad Advice

Today’s AI platforms are far from perfect. The cases listed above are just a handful of countless instances of generative and predictive AI platforms dispensing inaccurate information and providing bad advice. For professionals, this should be reason enough not to rely on machine learning to create work products or make judgments on their behalf. In no profession does reliance on AI currently serve as a complete defense against professional or malpractice liability.

AI “Hallucinations”

Beyond producing inaccurate information and rendering bad advice, AI platforms have also been known to create false information out of thin air. This is known as AI’s “hallucination” problem. AI’s propensity to hallucinate is yet another reason for professionals not to fully place their trust in generative or predictive AI platforms. Given that hallucinations are not always easy to spot, particularly when the subject matter is technical or unfamiliar, professionals need to be extremely cautious about both relying on AI for research purposes and representing AI-generated content as their own work product.

Implicit Bias

As the Harvard Business Review explains, “Human biases are well-documented, . . . [and o]ver the past few years, society has started to wrestle with just how much these human biases can make their way into artificial intelligence systems — with harmful results. At a time when many companies are looking to deploy AI systems across their operations, being acutely aware of those risks and working to reduce them is an urgent priority.” The implicit bias in AI algorithms is a significant concern for all types of professionals who use these algorithms in their practices. From dispensing financial or legal advice that reflects inherent bias to making biased treatment-related decisions in the healthcare setting, this is an issue across the professional spectrum.

Unintended and Unauthorized Disclosures

Overreliance on AI can also lead to unintended and unauthorized disclosures—and these can have professional liability implications in many cases. This includes not only unintended disclosure of confidential information (as in the venture capital firm case above), but also unauthorized disclosure of clients’ or patients’ personal information. The duty of confidence is among professionals’ most important duties in many professions, and breaching this duty can create exposure not only to professional liability, but to professional disciplinary action as well.

Other Errors, Omissions, and Oversights

Various other types of errors, omissions, and oversights are concerns when using AI in professional practice as well. From AI discovery tools failing to flag relevant documents in litigation to AI-powered medical device failing to identify complications during procedures, AI platforms can—and do—make many other types of mistakes as well. Once again, this simply underscores the importance of professionals not relying on AI exclusively, but rather using AI as one of many tools to assist with their decision-making in their professional practice.

Does (or Will) Professional Responsibility Ever Mandate the Use of AI technology?

Artificial intelligence platforms are improving every day, every hour, and every second. As they continue to improve, the risks discussed above will wane, and it will become more reasonable for professionals to place more trust in AI. This raises an important—and potentially transformative—question: Does (or will) professional responsibility ever mandate the use of AI system?

While the current answer is a pretty clear, “No,” this arguably has the potential to change over time. It may not change across the board, but it could change concerning certain aspects of professional practice. If an unbiased AI algorithm is capable of consistently producing better outcomes than professionals can produce on their own, it is feasible that it could constitute malpractice not to leverage the AI tools that are available.

Of course, from considering the costs involved to determining when relying on AI is definitively the best approach, there are many factors to consider—and, while this is a question worth asking, the fundamental nature of professional practice is not likely to change any time soon.

For now, professionals can certainly consider using AI tools when doing so helps to streamline their practice and enhance their productivity with AI risk management. However, it remains clear that overreliance on AI is a liability. Professionals who place too much trust in artificial intelligence tools can find themselves facing media liability, professional liability claims, cyber liability, and employment practices liability in AI litigation. Those who face AI-related malpractice allegations may be required to report these allegations to their licensing boards if patient harm happens. Given both the professional stakes and outcome-related risks involved, at this stage, overreliance on AI is a mistake that professionals in all sectors need to be very careful and address AI risks.

Schedule a Confidential Consultation with an Attorney at Oberheiden P.C.

If you have questions or concerns related to using AI in your professional practice, we invite you to get in touch. Our attorneys have extensive experience in this area, and we provide both consultation and representation for all types of AI-related legal matters. To schedule a confidential consultation at Oberheiden P.C., please call 888-680-1745 or inquire online today.

Further Information About Our AI Litigation Services

Why Clients Trust Oberheiden P.C.

  • 2,000+ Cases Won
  • Available Nights & Weekends
  • Experienced Trial Attorneys
  • Former Department of Justice Trial Attorney
  • Former Federal Prosecutors, U.S. Attorney’s Office
  • Former Agents from FBI, OIG, DEA
  • Serving Clients Nationwide
Contact Us 888-680-1745 866-781-9539