When does not using AI become negligence?

For the past several years, the legal conversation surrounding AI has largely focused on one question: What happens when professionals use AI and it gets something wrong? 

The headlines are familiar: lawyers sanctioned for briefs containing AI-generated hallucinations, biased algorithms influencing important decisions, copyright disputes over AI-generated content and confidential information exposed through AI tools. 

It is an important question and one that deserves continued attention. However, a different question is emerging as AI becomes more embedded in everyday work: What happens when professionals choose not to evaluate AI?

In conversations with clients, that concern is taking a more practical form — one that has less to do with the risks of using AI and more to do with the risks of not considering it. Increasingly, clients want to know what their peers are doing, whether AI has become the industry standard for a particular task and whether choosing not to use AI could place them at a competitive or even legal disadvantage. 

These questions reflect a broader concern. As AI becomes more commonplace, expectations of reasonable professional conduct may begin to change.

They go to the heart of a foundational concept of law: the standard of reasonable care. Courts have long evaluated professional conduct by asking what a reasonably prudent person or professional would have done under similar circumstances. But the standard of reasonableness has never been frozen in time and changes with the realities of professional practice.

We have seen this happen before. Technologies that began as optional innovations eventually became expected professional tools. The legal profession is no exception. 

There was a time when discovery meant reviewing and producing rooms full of paper documents. Today, electronic discovery is not simply more efficient, but a routine part of modern litigation. Legal research followed a similar path. What once required hours in a law library can now be accomplished in minutes using electronic research. 

Similar shifts have occurred across other professions. In healthcare, technologies like MRI and CT imaging have transformed the diagnostic process. In cybersecurity, tools like multifactor authentication have become basic expectations of a reasonable security system. 

Courts do not require professionals to adopt every new technology, but as some tools become established and widely accepted, they can influence what regulators and courts view as reasonable professional conduct. 

AI may be the next step in that evolution. Unlike many prior technologies, however, AI does more than make existing processes faster or more efficient. It increasingly assists with tasks that have traditionally required professional judgment — from analyzing information to drafting content and identifying issues for further review. 

As those capabilities become more reliable and widely adopted, the decision not to evaluate an available AI tool may itself become relevant to assessing whether reasonable care was exercised. 

That distinction matters because AI is increasingly being used not only to automate routine tasks, but to help professionals decide what deserves attention in the first place. An AI system may identify a contract provision for further review, flag a potential cybersecurity threat or identify patterns in medical images. The professional still makes the ultimate decision. 

Imagine a negligence lawsuit several years from now. A company suffers a significant loss after its legal team overlooks a critical contract provision that an AI review platform routinely identifies. During discovery, opposing counsel requests internal emails discussing AI adoption, governance committee minutes, procurement evaluations and analyses comparing the company's practices to those of its peers. 

The litigation would likely focus less on the use of AI and more on whether the organization's decision-making process reflected the level of care expected of a reasonably prudent company under similar circumstances.

For organizations, the practical challenge is establishing a guideline that helps professionals determine when AI is appropriate. That may mean creating a framework that allows professionals to evaluate available AI tools, understand their limitations, document their reasoning and exercise appropriate judgment. That documentation could become particularly important when a decision is later examined with the benefit of hindsight — after a failed transaction, a missed threat, an inaccurate analysis or another adverse outcome. Those governance decisions may eventually become evidence in litigation.

None of this means AI should be used simply because the technology exists. Reasonable judgment does not mean mandatory adoption. Just as courts do not expect professionals to use every available tool, they are unlikely to expect blind reliance on AI. 

Human judgment, expertise and oversight will remain essential. If standards evolve, they are likely to reward informed decision making, not automatic adoption. 

The law rarely changes overnight. Standards of care evolve gradually as technologies become accepted, industries adapt and courts evaluate new facts. AI is likely to follow that same path. 

Organizations do not need to precisely predict where the law will end up, but they should begin building governance processes that allow them to thoughtfully evaluate AI, document their decisions and demonstrate that those decisions were grounded in reasonable professional judgment.