A new Hepi–Taylor & Francis paper argues that artificial intelligence can accelerate translational research—helping convert laboratory discoveries into real‑world applications—while cautioning that ethical safeguards are essential to protect trust in academia. The report highlights AI’s potential to make research more discoverable, to link cross‑disciplinary findings, and to speed analysis of complex datasets, citing an example from the University of Warwick where AI helped police parse thousands of messages in gender‑based-violence investigations. Authors warn of risks including compromised data quality, opaque algorithms, bias, and the potential deskilling of early‑career researchers who may over‑rely on AI outputs. The paper proposes pairing AI use with explicit metacognitive training so users can identify model errors and preserve critical evaluation skills. Experts quoted in the report, including UCL’s Rose Luckin, advocate for metacognition training to turn AI from a deskilling risk into an upskilling opportunity; the paper frames governance, transparency and researcher training as prerequisites for responsible AI adoption in translational projects.