Why Human Judgment Must Stay at the Center of Medical AI

D

Dr. Martha Boeckenfeld

LinkedIn Author

AI Governance & Quantum Keynote Speaker | Board Director & Advisor | Human-Centric Futurist | I help boards & C-suites close the Governance Gap | Host, The Edge of Tomorrow | Ex-UBS · AXA

In a recent LinkedIn post, Dr. Martha Boeckenfeld discusses how medical AI might support care in parts of East Africa where specialists are far away and local health workers carry difficult decisions. Her post uses EPFL’s open-source Meditron model to make a broader argument: expanding access to clinical knowledge does not remove the need for human judgment, compassion and ethical responsibility. This article reports on Boeckenfeld’s post and examines the case she makes.

“The smartphone delivers clinical knowledge. The caregiver brings judgment, compassion, and the courage to act.”

AI can extend access to clinical knowledge

Boeckenfeld sets out a stark picture: a low-cost smartphone, a health worker without a medical degree and a patient more than 100 kilometers from the nearest specialist. In that context, she presents Meditron as a way to bring medical knowledge closer to the point of care. The model, developed by researchers at EPFL with collaborators, was trained on medical literature and clinical practice guidelines, according to the post’s cited research.

Her framing emphasizes the potential value of that access without claiming that an AI model can replace a clinician. Boeckenfeld points to the technical capacity of large models to process medical material, then contrasts it with the human work involved in responding to a patient. The distinction is central to her argument: information may travel through a device, but care still depends on the person using it.

The caregiver remains accountable

“Zero algorithms can look a frightened patient in the eye,” Boeckenfeld writes, describing a limit that is not measured by a benchmark. She argues that when technology informs a decision, a local caregiver still carries its ethical weight. In her view, the promise of medical AI depends on how it meets human judgment in real settings, not only on what a model can do in testing.

That point also raises practical questions about deployment: how health workers assess an AI system’s recommendations, what support they receive, and who is responsible when decisions are difficult. Boeckenfeld’s post does not resolve those questions, but its emphasis on responsibility highlights why they matter.

Recognition tied to a wider leadership argument

Boeckenfeld connects her remarks to being named Tech Influencer of the Year by the Influencer Magazine Awards, saying the recognition reinforces her concern about AI’s growing role in decision-making. She frames leadership in AI as a responsibility to protect human dignity and ethical judgment as systems become more capable.

“A machine can provide information. Only a human being can provide care.”

That closing line captures the post’s central distinction. Boeckenfeld is not dismissing technology’s potential; she is arguing that its value in medicine should be judged by whether it supports people responsible for patients. Her question to readers about moments when empathy mattered more than technology extends the discussion beyond Meditron to the broader challenge of deploying AI without losing sight of human needs.

📝 About This Content

This article is based on insights shared by Dr. Martha Boeckenfeld on LinkedIn.

📅 Originally posted on November 3, 2026 | View original post on LinkedIn →