Inside the Machine: Trust, Truth, and the Future of Knowledge in an AI World

# Inside the Machine: Trust, Truth, and the Future of Knowledge in an AI World > Bilva Chandra has worked on AI safety, ethics and governance both within the frontier labs and inside the institutions trying to govern or influence their progress. [Read on Substack](https://lawwhatsnext.substack.com/p/inside-the-machine-trust-truth-and) · 2026-04-13 · Law What's Next --- A quick reminder: Law://WhatsNext is our vehicle to explore through dialogue (or occasional reflection) how leading lawyers, educators and technologists are using emerging tech to evolve how we practice and administer legal services. No hype — just practical conversations. 🎙️ This week we sit down with Bilva Chandra — who has spent the bulk of her career working on AI safety, ethics, and governance at the places where it matters most. Bilva’s CV reads like a guided tour of the AI governance landscape: she’s worked at OpenAI, RAND, the US AI Safety Institute (now CAISI), and most recently Google DeepMind. She’s been inside the frontier labs building the technology and inside the institutions trying to govern or influence it’s progress — often thinking about the same problems from both sides of the table. One of the catalysts for our conversation emanated from her recent contribution to a Google DeepMind paper — Architecting Trust in Artificial Epistemic Agents — exploring what happens when AI systems become active participants in how knowledge is created and shared. But it’s Bilva’s broader career at the intersection of AI and society that makes this conversation so compelling: she’s someone who genuinely cares about getting this right, and isn’t afraid to say when she’s worried. Listen Now Available here or on Spotify, Apple Podcasts, or wherever you enjoy your podcasts. We cover a lot of ground — from the practical challenge of making AI systems reliable enough for enterprise adoption, to the deeper worry about what happens to human judgment when cognitive work is increasingly offloaded to machines. What emerges is a picture of someone who is both genuinely optimistic about what AI can unlock and deeply clear-eyed about the societal fault lines it’s accelerating. Bilva doesn’t treat AI risk as a theoretical exercise. She frames it as a human problem — one tangled up with polarisation, declining trust in institutions, and an information environment that was already broken before the first LLM shipped. Subscribe now Key References Connect with Bilva Chandra — on LinkedIn | Or by subscribing to Role Model, her new newsletter on AI and society. “Architecting Trust in Artificial Epistemic Agents“ — The Google DeepMind and Google Research paper (March 2026) led by Nahema Marchal that kicked off this conversation, exploring what it means for AI agents to be epistemically trustworthy. If you enjoyed this conversation, please share it with someone who you think would find it valuable — especially anyone grappling with how to govern AI responsibly in their organisation. And if you have a moment, rate the show and tell us what landed — it helps us to reach more people and attract brilliant guests like Bilva. We hope you have a great week! Tom & Alex Thanks for reading Law://WhatsNext! This post is public so feel free to share it. Share