In this episode of Data & AI Exchange (DAX) by Unolabs, we ask the big question: Can AI be trusted in healthcare, pharma, and beyond?Our guest, Dr. Thibault Géoui — a leading AI voice in biotech, an AI consultant, and host of the Tech & Drugs podcast — breaks down the challenges of building trust in AI: from drug discovery to diagnostics and even robotic surgery.What You’ll Learn: - Why “trust in AI” goes beyond accuracy and fairness - The hidden risks of bias in pharma and life sciences data - How AI is already cutting drug R&D timelines from 6 years to 8 months - Real examples where AI outperforms humans in diagnostics - Why benchmarks don’t always capture AI performance - Will we ever trust AI enough for autonomous surgeries or initial diagnostics?Whether you’re an AI enthusiast, healthcare professional, or just curious about the future of technology, this conversation will give you real-world insights into where AI delivers, where it fails, and why trust is the ultimate barrier.Don’t forget to subscribe for more discussions on AI, trust, and the future of technology.00:00 – Can AI replace doctors in the next 10 years?00:29 – Guest intro: Dr. Thibault Géoui, biotech & AI leader01:12 – Why pharma is slow to adopt new technologies02:26 – The gap: 24,000 diseases vs. only 3,000–4,000 drugs03:33 – Drug development costs ($2–6B, 10–12 years) & 95% failure rate04:47 – Why pharma is excited about AI as a new tool05:50 – Trust in AI: accuracy, fairness, reliability, or full lifecycle?07:40 – Pharma’s AI adoption “purgatory” explained09:01 – Why current AI benchmarks don’t tell the full story11:44 – Can bias in AI models ever be fully eliminated?12:53 – Pharma’s own bias: drugs developed for a “typical” patient14:40 – Biological diversity in drug effectiveness (genetics, enzymes, etc.)16:26 – Data sourcing challenges: pharma sitting on tons of unusable data18:07 – Historical vs. new data: accessibility and usability19:17 – Vinod Khosla’s prediction: AI watch for early diagnostics20:15 – AI in cancer imaging: from manual to automated precision22:11 – Specific AI use cases outperforming humans (chess, Go, diagnostics)23:01 – Cutting drug R&D timelines: from 6 years to 8 months24:40 – Preparing oversight mechanisms for AGI-level AI25:16 – Lessons from self-driving cars & autopilot in planes26:20 – Robotic arms in surgery: augmentation vs. full automation27:55 – Sci-fi to reality: Clarke’s novel vs. today’s surgical robots29:07 – AI-assisted vs. autonomous surgery — what’s realistic?29:48 – Closing thoughts: AI’s future in healthcare & pharma#CanAIBeTrusted #ArtificialIntelligence #AIinHealthcare #AItrust #PharmaInnovation #DrugDiscovery #AIMedicine