AI in Healthcare

Today's Stories

1. 1

   Google's medical AI matches real doctors at managing chronic illness, study finds

   Google published peer-reviewed research in Nature on June 17 showing that AMIE (its conversational AI system designed for medical consultations) performed as well as primary care physicians when managing patients with complex, ongoing conditions like diabetes or heart disease. The study involved simulated patient conversations evaluated by independent clinicians. This is the first time a large-scale study in a top scientific journal has validated an AI holding its own against trained doctors in multi-turn disease management — not just answering one-off medical questions.

   If this holds up in real clinical trials, AI-powered chat tools could help fill gaps in primary care access — meaning people in underserved areas or on long waiting lists could get reliable medical guidance without waiting months for a doctor's appointment.

2. 2

   Merck bets up to $510 million that AI startup Protillion can find new drugs faster

   Drug maker Merck announced on June 16–17 a collaboration with Protillion, an AI drug-discovery startup, worth up to $510 million in milestone payments (bonuses paid out as research hits specific targets). Protillion uses AI to identify promising drug candidates — molecules that might treat disease — far more quickly than traditional lab-based trial-and-error methods. Deals like this typically start with an upfront payment plus large bonuses if the AI-discovered drug passes clinical trials.

   When major pharmaceutical companies invest this heavily in AI-driven drug discovery, it accelerates timelines for new medicines reaching patients — potentially cutting years off the decade-long process it normally takes to bring a drug to market.

3. 3

   Off-the-shelf AI like ChatGPT outperforms purpose-built hospital AI, Nature study reveals

   A study published in Nature Medicine on June 16 found that general-purpose large language models (LLMs — AI systems trained to understand and generate text, like those powering ChatGPT) performed better on clinical tasks than specialized AI tools built specifically for healthcare settings. This challenges the assumption that narrowly trained, hospital-specific AI is worth the premium cost. The finding has significant implications for hospitals and health systems that have spent heavily on dedicated clinical AI software.

   Hospitals and clinics spending large budgets on expensive specialized AI software may get equally good — or better — results by simply using widely available general AI tools, which could free up healthcare dollars for direct patient care.

4. 4

   Pfizer licenses AI drug design technology from startup Chai

   Pfizer signed a licensing agreement with Chai, an AI startup that specializes in predicting how molecules (the building blocks of drugs) fold and interact — a critical step in designing new medicines. The deal, announced June 9, gives Pfizer access to Chai's AI platform for its drug-discovery pipeline. Chai's technology is similar in concept to DeepMind's AlphaFold, which revolutionized biology by predicting protein structures, but focused specifically on drug design applications.

   Pfizer adding AI molecular design to its toolkit means future drugs — including vaccines and cancer treatments — could be engineered more precisely, potentially making them more effective with fewer side effects.

5. 5

   Startup Radical Numerics raises $50M to generate entirely new DNA sequences with AI

   Radical Numerics, founded by researchers who pioneered AI-generated DNA, closed a $50 million funding round on June 15. The company uses AI to design novel biological sequences — essentially writing new genetic code from scratch — for applications in drug discovery and biodefense (defending against biological threats). Unlike approaches that study individual molecules, Radical Numerics models entire biological systems at once, which its founders argue gives a fuller picture of how a disease works and how to fight it.

   This technology could eventually lead to faster development of vaccines or treatments for emerging diseases and biological threats — compressing timelines from years to months when the next pandemic or outbreak occurs.

6. 6

   New AI startup Pramaana Labs raises $27M to make AI give reliable answers in high-stakes fields

   Pramaana Labs secured a $27 million seed round (early-stage startup funding) from Khosla Ventures on June 17 to apply 'formal verification' (a mathematical method that proves software behaves correctly, the same technique used in aviation and nuclear systems) to AI outputs. The company will focus first on drug discovery, legal work, and tax preparation — fields where an AI making a confident but wrong answer can have serious real-world consequences. Current AI systems can produce plausible-sounding errors, a problem the industry calls 'hallucination.'

   If Pramaana's approach works, it would make AI-generated medical or legal advice genuinely trustworthy — rather than requiring experts to double-check every AI output, which currently limits how much time AI actually saves in professional settings.

7. 7

   Investors say the next big healthcare AI money-makers are drugmakers and biotech firms, not tech giants

   A June 19 analysis by The Motley Fool argues that the next wave of AI-driven wealth creation in healthcare will flow to pharmaceutical and biotech companies — such as Eli Lilly (known for its weight-loss drug Mounjaro) and Twist Bioscience (which manufactures synthetic DNA for research) — rather than to the AI technology companies themselves. The thesis: the real value of AI in medicine is captured by the organizations that own the drugs and biological platforms AI helps create, not the AI tool vendors.

   For anyone with retirement savings or investment portfolios, this framing suggests looking beyond obvious AI chip and software stocks toward life-sciences companies actively integrating AI into their research pipelines.