AI Tools in Medicine: Legal Gaps and Liability Questions

As new technologies change core aspects of modern healthcare, the field of medicine has been racing to keep pace with the rapid growth of artificial intelligence (AI). Applications in the medical field include diagnostics, treatment planning, imaging analysis, and patient monitoring. Though gaining traction in clinical settings, AI isn’t error-free, which begs the question: Who is legally responsible when AI makes a mistake?

For injured patients filing a negligence claim, their Ohio medical malpractice lawyer must understand how these emerging issues can substantially increase uncertainty regarding fault.

In this article, we take a magnifying glass to the key liability risks and gaps in AI-assisted medical care, and how these risks and gaps intersect with traditional medical malpractice law.

How is AI used in medicine?

AI is used by doctors, nurses, and other healthcare professionals for a variety of tasks:

• Machine learning to read medical images
• Predictive algorithms for assessing patient risk
• Decision-support systems for recommending treatments
• Monitoring systems for identifying physiological changes

While helpful in increasing efficiencies, these tools typically use vast datasets and intricate algorithms. Often, since AI models function as “black boxes,” their internal logic is either inexplicable or unclear overall. Errors can stem from flawed training data, biased data, or performance degradation over time.

Where does liability risk arise?

As AI is progressively used in the medical field and courts hear more of these cases, legal doctrine will be compelled to evolve. Currently, the boundaries between physician error, institutional oversight, and software defect are blurred.

Though the lines of liability are blurred, there are a few distinct situations where liability is likely to arise:

• Medical malpractice – if providers negligently rely on AI
• Product liability – if the AI system is defective
• Negligent deployment or oversight – failing to update or monitor tools
• Regulatory noncompliance – breaching FDA rules or privacy laws ( but there is no private right of action under the FDCA)

Legal uncertainties in AI-assisted medicine

Unlike conventional medical malpractice claims, AI-assisted medical care introduces new actors and a web of causal chains. Legal standards have yet to catch up with the inclusion of AI in healthcare, so uncertainties now make it hard to predict case outcomes and increase the burden on plaintiffs.

Who is the bad actor?

If AI is just a tool, like a stethoscope, the physician will probably be held liable. If AI is considered more autonomous, liability may shift toward the healthcare institution or software company.

As AI is still emerging (and not fully trusted), many courts are holding healthcare professionals and facilities liable. Doctors are expected to closely evaluate AI output and not take it at face value.

Burden of proof and expert testimony

To prove medical malpractice, the plaintiff must show:

1. Breach of the established standard of care
2. Causation of harm
3. Resulting damages

Expert testimony encompasses medical and technical experts who can assess how the AI system was trained, validated, and used in practice.

Allocation of fault among defendants

AI-related harm often involves multiple contributors, such as physicians, hospitals, and software companies. Under Ohio’s modified comparative fault, a plaintiff who is more than 50% at fault is barred from recovery. Defendants are severally liable for noneconomic damages in proportion to fault, but a defendant >50% at fault (or an intentional tortfeasor) is jointly and severally liable for all economic damages.

In an AI case, liability may be allocated across the:

• Doctor who trusted a flawed output;
• Hospital that didn’t update the AI system; and/or
• The vendor that released a defective product.

A thorough investigation and expert analysis must be conducted to parse through the various contributions.

Evolution of the standard of care and AI adoption

As AI tools become more widespread and dependable, the medical standard of care may shift. In the future, failure to use certain AI tools in a medical practice may be viewed as negligent, although early adoption carries significant risk. Ohio medical malpractice lawyers would be prudent to monitor how courts treat these transitional cases.

Software updates, versions, and drift

AI models can degrade over time due to changes in patient data, known as data drift. If a hospital or institution doesn’t revalidate a system after a software update or keeps using an outdated version, it may be held responsible for not keeping it up to date.

Tracking the version of the AI tool used at the time of harm, performance logs, and audit trails is crucial in terms of proving causation.

What you can and can’t recover in Ohio

Ohio caps noneconomic damages in medical claims at the greater of $250,000 or three times the economic loss, up to $350,000 per plaintiff and $500,000 per occurrence. The cap does not apply to wrongful death actions. Recent appellate decisions have found the cap unconstitutional as applied in some catastrophic‑injury medical cases.

For qualifying injuries defined in R.C. 2323.43(A)(3) (e.g., permanent and substantial physical deformity; loss of use of a limb; loss of a bodily organ system; or permanent physical functional injury preventing independent self‑care and life‑sustaining activities), the noneconomic cap increases to $500,000 per plaintiff / $1,000,000 per occurrence.

On the other hand, the state doesn’t place limitations on economic damages like medical expenses, lost income, or future care costs. An Ohio medical malpractice lawyer can help you understand your options and maximize your recovery.

Tips for handling negligence cases involving AI

Best practices for cases based on AI tools in medicine aren’t too different from standard medical negligence cases.

Early expert involvement

Make sure medical and technical experts, including those with AI experience, are involved from the start. These experts can assess and explain system design, validation, and outputs, as well as determine whether the outputs meet the standard of care.

Preserve evidence

Request logs, software versions, model training data, and audit trails from vendors and/or hospitals. These are critical pieces of information and documentation that can be used to establish causation or fault.

Keep a watchful eye on the statute of limitations

Unfortunately, complex issues don’t extend the statute of limitations. Ohio’s strict standard malpractice deadlines still apply for AI-related cases.

Map fault allocations at the outset

It’s important to be strategic when identifying defendants and eventually assigning fault. AI cases commonly involve multiple contributors, so it’s important to be mindful of Ohio’s comparative fault law.

What to do if AI tools used in your medical care have failed you

The integration of AI in medicine holds tremendous promise and legal uncertainty, too. For Ohio patients who’ve been injured because of AI-assisted medical errors, bridging this gap between emerging technology and old medical malpractice laws can be overwhelming.

Our team of lawyers knows how complicated these cases can be from a medical and technical point of view. At Crandall & Pera Law, we’re ready to review AI evidence, coordinate expert analysis, and explore liability against all potentially responsible parties. Schedule a consultation to start the process of protecting your rights as AI continues to reshape healthcare as we know it.