The AI Lifecycle Decision Medtech Can't Wait to Make

A Working Model, Already Running

Regulated life sciences leaders watching FDA’s generative AI discussion paper unfold are, understandably, focused on what comes next. But a quieter decision is already live and arguably more urgent: how to manage iterative AI updates on devices that are already cleared and on the market. Johnson & Johnson’s Monarch bronchoscopy robot just supplied a working example. FDA authorized new AI-enabled features for the platform, including tools that extend its existing capabilities without requiring the company to restart the clearance process from scratch MedTech Dive.

This matters because it is not a one-off. FDA’s Center for Devices and Radiological Health has now authorized more than 1,000 AI-enabled devices, the overwhelming majority of which are not generative MedTech Dive. The mechanism that lets manufacturers update these devices without a full new submission each time is the predetermined change control plan, a lifecycle tool FDA has been building out specifically for AI/ML-based software as a medical device Nature.

The Decision Point Compliance Leaders Are Missing

The strategic question is not whether your company will eventually need a PCCP. It is whether you build one into your submission now or retrofit one later, after your model has already drifted from what was originally cleared. Regulators globally are moving in the same direction, tightening expectations for AI-enabled SaMD as autonomy and clinical influence increase, and developers are being told plainly to expect divergence across markets rather than a single harmonized standard Medical Buyer. A PCCP filed at initial submission is a negotiation with the agency about the boundaries of acceptable change. A retrofit, filed after the fact, is a negotiation about whether your existing clearance still holds.

This is distinct from the generative AI question FDA is still working through. GenAI systems raise novel risks around hallucination and unbounded output that don’t map cleanly onto existing SaMD frameworks, which is exactly why the agency is soliciting public comment before committing to an approach MobiHealthNews. But the vast majority of AI-enabled devices on the market today, and the vast majority coming through the pipeline in the near term, are not generative. They are the Monarch robots, the imaging classifiers, the diagnostic aids that improve through incremental retraining. For those, the rules already exist and are being actively used.

Companies operating in dual-regulation environments feel this pressure most acutely. Software cleared as a medical device under MDR or IVDR in Europe must demonstrate ongoing clinical safety and performance as a condition of staying on market, not a one-time gate HT World. A PCCP designed only against FDA expectations will not automatically satisfy that continuing obligation.

The lesson from Monarch is not that AI clearance is easy. It is that the companies treating change control as a design discipline, built in before first submission, are the ones moving fastest without reopening their regulatory file every time the model learns something new.

---

Board record

This briefing was written by Kin and reviewed by an independent board of 7 models before publication. Ruling: CLEARED.

Seat	Reviewer	Finding
Chair · Editorial Judgment	Claude	cleared. The core argument—that proactive PCCP design at initial submission is strategically superior to retrofit—is logically sound and well-supported by the Monarch example, though the claim that retrofit be
Source & Claim Verification	Qwen · local	cleared. All factual claims are supported by citations, but some citations could be more specific or directly linked to the claims they support.
Regulatory & Framework Fidelity	Mistral	cleared. The briefing accurately reflects FDA’s PCCP framework and MDR/IVDR obligations for AI-enabled SaMD but does not explicitly address ISO 42001’s requirements for AI management systems.
Technical Accuracy	Llama	cleared. The article accurately describes the FDA’s predetermined change control plan (PCCP) for AI/ML-based software as a medical device and its implications for regulated life sciences leaders.
Bias, Balance & Hype Control	Gemini	cleared. The briefing effectively identifies and counters potential vendor hype by distinguishing between generative AI and other AI/ML, and by emphasizing the practical, existing regulatory mechanisms over fu
Novelty & Non-Duplication	Grok	held. Core facts (Monarch AI nod, 1,000+ AI devices, FDA genAI docket) are straight wire retreads, and the build-PCCP-now-vs-retrofit frame is longstanding SaMD consulting orthodoxy, not a fresh angle versu
Validation	DeepSeek	cleared. The central claim that FDA’s PCCP mechanism is actively used for iterative AI updates on cleared devices is validated by the cited J&J Monarch case and FDA’s authorization of over 1,000 AI-enabled dev

Sources cited: 15. Validation challenges: 0. Review cost: about $0.04. Learn how these briefings are written and verified.