Agentic AI

Designing human oversight for clinical agents

How to build supervision around an AI agent that acts on the clinical record — translating EU AI Act Article 14 and FDA guidance into concrete oversight tiers, automation-bias defences, and after-go-live monitoring. As of July 2026.

By Jonas WeirReviewed by Jonas Weir · editorial reviewUpdated

The short version

  • Human oversight of a clinical agent is a design problem rather than a checkbox: EU AI Act Article 14 requires high-risk systems be built so a person can understand, interpret, override, and stop them — and remain aware of automation bias.
  • Match the oversight grade to the consequence of the action: in-the-loop approval for anything that writes to the record, on-the-loop monitoring for reversible low-risk steps, and a hard stop control for every tier.
  • A human being present differs from a human effectively checking: a systematic review of 74 studies documented automation bias across decision-support settings, so oversight must be engineered against rubber-stamping.
  • US FDA excludes decision-support software from device regulation only when the clinician can independently review the basis for a recommendation — an agent that acts, rather than advises, strains that line.
  • Oversight continues after go-live: transparency requirements for predictive decision support and FDA lifecycle guidance both point to logging, source disclosure, and monitoring as standing obligations. As of July 2026.

An agent that only drafts is easy to supervise: a clinician reads the draft and signs. An agent that can act on the record — place an order, send a message, close a task — changes the question from is the output correct to who is accountable for what it did. This page is about building the answer: how to design human oversight around a clinical AI agent, translating the two anchors that govern it — EU AI Act Article 14 and US FDA guidance — into oversight tiers, automation-bias defences, and monitoring that continues after go-live. It is the applied companion to the human-in-the-loop definition. As of July 2026.

What oversight legally requires

Two regimes set the floor, and they agree on the direction.

The EU AI Act requires that high-risk AI systems — a category that captures most consequential clinical uses — be designed so they "can be effectively overseen by natural persons during the period in which they are in use" 1. Article 14(4) then enumerates what that oversight person must be enabled to do. Read as a design spec, it is a list of capabilities the system has to provide:

Art. 14(4)The overseer must be able toDesign implication for an agent
(a)Understand the system's capacities and limitations; monitor its operation and spot anomaliesExpose what the agent is doing and surface unexpected behaviour
(b)Remain aware of automation bias — over-relying on the outputInterface and training that counter reflexive acceptance
(c)Correctly interpret the outputShow the basis, the inputs, and the confidence behind an action
(d)Decide not to use, or disregard, override, or reverse the outputAn override path that is real, logged, and reversible
(e)Intervene or interrupt the system via a stop controlA working stop button at every tier of autonomy

In the United States, FDA draws the line at a different place but points the same way. It excludes a clinical decision support system from device regulation only when, among other criteria, the software enables the clinician to independently review the basis for its recommendations rather than rely on them primarily 2. The logic matters for agents: a system that advises a reviewer who retains judgement can qualify for the exclusion, while a system that acts — where the human reviews after the fact, if at all — strains it, and more readily falls under device oversight. WHO's guidance on large multi-modal models reinforces the principle, cautioning against deployments that leave model output unchecked 5.

Match the oversight grade to the action

The human-in-the-loop literature distinguishes two postures. In-the-loop: a person acts on every output — the ambient-scribe pattern, where a clinician edits and signs each draft before it enters the record. On-the-loop: a person monitors a running system and intervenes on exception. The design mistake is to pick one posture for the whole agent. The safer approach is to grade oversight by the consequence of each action the agent can take.

Action classExampleOversight gradeRationale
Read-onlyRetrieve labs, summarise a chartOn-the-loop, loggedReversible, low harm
Draft-for-reviewCompose a note, propose a planIn-the-loop sign-offHuman judgement before the record changes
Write / orderFile an order, send a patient messageIn-the-loop approval, hard gateConsequential and hard to reverse
EscalationLow confidence, conflicting data, unsafe requestRoute to a human, agent stopsFailure and edge cases belong with a person

This grading follows directly from where the risk lives. As agentic AI benchmarks show, capable systems still fail a meaningful share of tasks — the best model on MedAgentBench completed 69.67% of realistic EHR tasks, leaving roughly three in ten to fail 6. Oversight cost should be spent where a failure is both likely and costly: the write and order actions. Spending equal scrutiny on a read-only lab lookup wastes the clinician attention that the consequential actions need.

Blanket in-the-loop review also carries a cost that shows up as disuse. In a roughly 1,800-clinician study of ambient documentation, where every AI-drafted note required review before signing, daily documentation time fell — yet only 32% of users adopted the tool for more than half their visits 9. Oversight is real clinical work, and a design that demands equal effort at every step spends the reviewer's attention thin. Grading actions by consequence is what keeps that scarce attention on the decisions that carry the harm. As of July 2026, this trade-off — between the safety the loop adds and the effort it costs — is the central tension in oversight design.

Designing against automation bias

The hardest problem in oversight is that a human being present differs from a human effectively checking. The EU AI Act names the failure mode in the text itself: overseers must "remain aware of the possible tendency of automatically relying or over-relying on the output" — automation bias 1. That clause exists because the failure is well-documented. A systematic review of 74 studies recorded automation bias across clinical decision-support settings, and identified its mitigators: training, accountability, and interface design 4. WHO repeats the warning for large multi-modal models, noting they may encourage automation bias whereby errors are overlooked that would otherwise have been caught 5.

Designing against it means engineering the review, rather than trusting it. Concretely:

  • Make the basis prominent. An overseer who can see the inputs, the retrieved evidence, and a confidence signal can interpret an action, as Article 14(4)(c) requires 1. An action presented as a bare recommendation invites reflexive acceptance.
  • Design for the error, not the average. Review works best when the reviewer expects to find something. Surfacing the agent's uncertainty, and flagging the cases it found hardest, directs attention toward the outputs most likely to be wrong — the hallucination and omission failures review exists to catch.
  • Protect the reviewer's capacity. Oversight is work, and a reviewer under volume pressure rubber-stamps. Grading actions by consequence keeps scarce attention on the decisions that carry it, rather than spreading it thin across every step.
  • Locate accountability. Article 14(4)(d) requires a real path to disregard, override, or reverse an output 1; accountability for having done so — a named owner, a logged decision — is what makes that path more than nominal.

A worked example shows how these combine. An agent proposes a medication order. A design that counters automation bias presents the order beside the inputs that drove it, the guideline it retrieved, and a confidence signal; it defaults to requiring an explicit approval action rather than a passive time-out; it flags that this case sat in the low-confidence band the model found hard; and it records who approved it and when. The same order presented as a one-line recommendation with an accept button invites exactly the reflexive acceptance the review of 74 studies warns about 4. The interface is the oversight, as much as the person behind it.

An oversight pattern for agents

Bringing the pieces together, a workable pattern for a clinical agent has five standing elements.

  1. Action tiering. Classify every tool the agent can call by consequence, and attach the oversight grade from the table above. This is the backbone; it turns "a human is involved" into "this specific action requires this specific approval."
  2. A hard stop at every tier. Article 14(4)(e) requires the ability to interrupt or stop the system 1. The stop control must halt an in-progress action, rather than merely preventing the next one.
  3. Basis-on-display. Every consequential action shows its inputs, evidence, and confidence, so the overseer can satisfy 14(4)(c) rather than approve blind.
  4. Logging and transparency. Under the US predictive decision-support transparency requirements, certified health IT must disclose source attributes for predictive interventions so users can judge whether they are fair, appropriate, valid, effective, and safe 7. Applied to an agent, the principle is a durable audit trail: what it did, on what basis, and who approved it.
  5. An escalation route. Low confidence, conflicting data, or an unsafe request routes to a human and the agent stops. The scoping review's three archetypes — conversational, workflow, and decision-support agents 8 — each need their own escalation triggers, because a triage conversation fails differently from an order-drafting workflow.

Oversight after go-live

Oversight is a standing function, rather than a launch gate. FDA's 2025 draft guidance on AI-enabled device software functions frames expectations across the total product life cycle, including monitoring performance after deployment 3. Agents drift as their models, prompts, tools, and patient populations change, and an oversight design that is never revisited decays with them. The monitoring side of this belongs to algorithmovigilance — the ongoing surveillance of a deployed model's behaviour — and it should answer standing questions: what is logged, which metrics are watched, who owns the watch, and what triggers a pause or rollback. The regulatory expectations themselves keep moving; we track them in the global AI in health regulation tracker.

Two monitoring signals deserve naming because an agent produces them and a static model does not. The first is the intervention rate: how often overseers disregard, override, or reverse the agent's proposed actions, broken down by action tier. A rising override rate on write actions is an early warning that the system's judgment and the clinicians' have drifted apart. The second is the escalation rate: how often the agent stops and hands off. Both belong in the audit trail alongside the tool calls themselves, so that a later review can reconstruct what the agent did, on what basis, and who approved it — the record the transparency requirements are built to support 7. Where several agents collaborate, that trail has to span them, since a failure can originate in one role and surface in another.

How to read this

Three cautions travel with the pattern. First, regulation here is live: the EU AI Act's Article 14 obligations are in force for high-risk systems, but harmonised standards and the FDA lifecycle guidance are still maturing as of July 2026, so confirm the current requirements with your compliance or regulatory counsel before acting. Second, oversight design is contextual — the right tiering for an order-drafting agent differs from that for a patient-messaging one, and the tables here are a starting frame to adapt, rather than a fixed rule. Third, oversight complements a safe system; it does not substitute for one. Review catches errors imperfectly, and automation bias limits how much it can catch 4, so the control works best when the agent's own error rate is already low — which is what pre-deployment evaluation is for.

Sources and method

This guide translates two governing anchors — EU AI Act Article 14 1 and FDA guidance on clinical decision support 2 and AI-enabled device lifecycle management 3 — into oversight design, grounded in the automation-bias literature 4, WHO guidance on large multi-modal models 5, the transparency requirements for predictive decision support 7, and the benchmark and scoping evidence that shows why oversight is necessary and where failures cluster 68. Every figure and requirement is drawn from the primary source cited beside it. We revisit this page on a 180-day cycle and whenever a governing rule is revised. For the concept, see human-in-the-loop; for the system these controls wrap, see what is agentic AI in healthcare.

Questions & answers

  • Is human oversight legally required for a clinical AI agent?

    For high-risk uses, yes — as a design requirement. The EU AI Act requires high-risk AI systems to be built so natural persons can effectively oversee them, and US FDA excludes decision-support software from device regulation only when the clinician can independently review the basis for its recommendations. An agent that acts on the record, rather than advising a reviewer, tends to fall on the regulated side of that line.

  • What is the difference between in-the-loop and on-the-loop oversight?

    In-the-loop means a person approves each individual output before it takes effect. On-the-loop means a person monitors a running system and intervenes on exception. For a clinical agent, the safe default is to match the grade to the action: in-the-loop approval for anything that writes to the record, on-the-loop for reversible low-risk steps, with a stop control at every tier.

  • Does putting a human in the loop guarantee safety?

    No. Automation bias — the documented tendency to over-rely on automated advice — means a nominal reviewer can wave through errors they would have caught unaided. The EU AI Act names this failure directly and requires overseers to stay aware of it. Effective oversight has to be engineered through interface design, workload limits, and accountability, rather than assumed from a person's presence.

Sources

  1. Regulation (EU) 2024/1689 of the European Parliament and of the Council (Artificial Intelligence Act), Article 14: Human oversight. Official Journal of the European Union. 12 July 2024. eur-lex.europa.eu/eli/reg/2024/1689/oj
  2. US Food and Drug Administration. Clinical Decision Support Software — Guidance for Industry and FDA Staff. Revised final guidance, January 2026. www.fda.gov/regulatory-information/search-fda-guidance-documents/clinical-decision-support-software
  3. US Food and Drug Administration. Artificial Intelligence-Enabled Device Software Functions: Lifecycle Management and Marketing Submission Recommendations; Draft Guidance. Federal Register. 7 January 2025 (FDA-2024-D-4488). www.federalregister.gov/documents/2025/01/07/2024-31543/artificial-intelligence-enabled-device-software-functions-lifecycle-management-and-marketing
  4. Goddard K, Roudsari A, Wyatt JC. Automation bias: a systematic review of frequency, effect mediators, and mitigators. Journal of the American Medical Informatics Association. 2012;19(1):121-127. doi.org/10.1136/amiajnl-2011-000089
  5. World Health Organization. Ethics and governance of artificial intelligence for health: Guidance on large multi-modal models. Geneva: WHO; 18 January 2024. www.who.int/publications/i/item/9789240084759
  6. Jiang Y, Black KC, Geng G, et al. MedAgentBench: A Virtual EHR Environment to Benchmark Medical LLM Agents. NEJM AI. 2025. doi.org/10.1056/AIdbp2500144
  7. Health Data, Technology, and Interoperability: Certification Program Updates, Algorithm Transparency, and Information Sharing (HTI-1 Final Rule). Federal Register. 9 January 2024. www.federalregister.gov/documents/2024/01/09/2023-28857/health-data-technology-and-interoperability-certification-program-updates-algorithm-transparency-and
  8. Njei B, Al-Ajlouni YA, Kanmounye US, et al. Artificial intelligence agents in healthcare research: A scoping review. PLOS One. 2026;21(2):e0342182. doi.org/10.1371/journal.pone.0342182
  9. Changes in Clinician Time Expenditure and Visit Quantity With Adoption of Artificial Intelligence–Powered Scribes. JAMA. 2026. doi.org/10.1001/jama.2026.2253