Press room / current record

Podcast appearance · AI Ethics Now · University of Warwick IATL

AI and Accountability: When Does a Suggestion Become a Decision?

Varun Sharma joins AI Ethics Now to examine the line between an AI-assisted suggestion and a decision for which a human institution must remain accountable.

Varun SharmaGuest · 23 August 2026

Illustrative high-end podcast studio with microphones, headphones and a mixing console
Illustrative editorial image · no recording venue or endorsement implied
Read this record accurately.

A podcast appearance is evidence of a published conversation. It does not imply a University of Warwick partnership or endorsement, and it is not an assessment of PHARMORIS, its maturity or its readiness for use.

The full conversation, in context.

The Spotify player stays blocked until you choose to load it. Spotify may then process technical data or place cookies under its own terms.

Open the episode on Spotify
01

When assistance becomes authority

The useful governance question is not whether AI was present. It is when a suggestion starts to shape a consequential choice, who can challenge it and who remains answerable when the system sounds more certain than the evidence permits.

02

Human oversight must be operational

A human name on a diagram is not enough. Responsible systems need an identifiable decision owner, visible source limitations and a real ability to accept, reject, defer or request further evidence without losing the rationale.

03

Medicines expose the stakes

Supply disruption and medicines operations make accountability tangible: a recommendation can travel quickly while authority, local context and the consequences of acting remain distributed across real people and institutions.

The facts, without the halo effect.

What is AI Ethics Now?

AI Ethics Now is a University of Warwick Institute for Advanced Teaching and Learning podcast. This record concerns one published episode featuring Varun Sharma.

What did Varun discuss?

The conversation examines false confidence, identifiable human responsibility, medicines disruption, institutional adoption and the difference between automation and domain judgement.

Does the appearance validate PHARMORIS?

No. Participation in the podcast does not establish product readiness, deployment, clinical validation, customer adoption or endorsement by the University of Warwick.

Return to the complete publication record