The pharmaceutical industry has largely solved the challenge of adopting artificial intelligence, but a harder problem is now taking centre stage: how to properly supervise it.
Amy Brown, founder and CEO of Authenticx and a former healthcare executive, is making the case that continuous oversight of patient-facing AI systems is the sector’s most pressing priority right now.
Most pharmaceutical companies still rely on manual sampling and general performance metrics to evaluate whether their AI systems are functioning as intended.
Brown argues this approach is insufficient for regulated patient interactions, which demand a far more rigorous and healthcare-specific standard of evaluation.
The concern is not simply about accuracy or efficiency, but about whether AI deployed in patient access and support operations is behaving appropriately across every interaction.
Brown notes that over the past two years there has been an explosion in AI-enabled tools, with pharma adoption moving fastest in the use of frontier models like ChatGPT and Claude.
These tools have been deployed primarily to help the workforce manage everyday tasks, representing an early and relatively straightforward phase of integration.
The more complex challenge now involves supervising those systems in real time, particularly where clinical and regulatory judgment must become explicit evaluation criteria.
Brown’s position is that AI monitoring AI, rather than human teams conducting periodic spot checks, represents the practical model for achieving meaningful oversight at scale.
When clinical knowledge and regulatory requirements are embedded into evaluation frameworks, AI systems can be assessed against the standards that actually govern patient interactions.
Without this approach, pharmaceutical companies risk operating AI in a compliance blind spot, where systems may perform well on generic benchmarks but fail on healthcare-specific criteria.
The shift Brown describes is significant because it reframes the industry’s relationship with AI from implementation as a milestone to supervision as an ongoing operational responsibility.