Overview of Rachman Review
This episode of the Rachman Review (Financial Times) features Gideon Rachman’s interview with Bill Gates in Davos. The conversation centers on global development (especially recent setbacks in child mortality and international aid), Gates’s philanthropic capacity and limits, and the Gates Foundation’s pilot AI-healthcare partnership with OpenAI (Horizon 1000). The interview also touches on Gates’s reactions to the Epstein files, his dealings with Donald Trump, and broader risks and benefits of AI.
Key topics discussed
- Recent rise in under‑5 child deaths and causes (2024 → 2025 uptick).
- Global aid cuts (USAID and several sovereign donors) and the effects on vaccine and bed‑net distribution.
- Limits of philanthropy vs. sovereign aid and Gates Foundation’s 20‑year spending plan.
- Horizon 1000: Gates Foundation + OpenAI pilot to deploy AI in 1,000 African clinics (starting Rwanda).
- Technical details and constraints of AI healthcare deployment (languages, connectivity, interfaces).
- AI benefits and risks (medical gains, labor disruption, bioterrorism, misinformation).
- Gates’s view on public controversies: Epstein associations and engagement with President Trump.
Main takeaways
- Child‑mortality progress reversed slightly: after a historic fall from >10 million annual under‑5 deaths (circa 2000) to ~4.6 million (2024), Gates reports an increase to ~4.8 million in 2025. Disruptions to primary healthcare—linked to donor cuts—are a major cause.
- Sovereign aid remains far larger than philanthropic giving: U.S. government aid is ~US$40bn/year (humanitarian/delivery), the U.K. was previously comparable, while Gates expects foundation spending to reach up to ~US$12bn/year at peak. Philanthropy cannot fully replace government aid.
- Gates and the foundation will spend essentially all his personal wealth and foundation resources over the next 20 years to accelerate progress, but that will not compensate for large sovereign cuts.
- Horizon 1000: a US$50m mutual commitment (OpenAI + Gates Foundation) to pilot AI tools in 1,000 African clinics (Rwanda first) to assist clinical record‑keeping, symptom evaluation and decision support—especially for nurse‑practitioners and common conditions (HIV, maternal care).
- Practical deployment constraints: local language/dialect data collection (benchmarking ~40 African dialects), connectivity needs (voice vs. data/smartphone), and the preference to pilot where a data/smartphone interface is available.
- Gates is optimistic about AI’s ability to amplify medical knowledge and productivity but warns of substantial risks: misuse by bad actors (notably bioterrorism), misinformation, and rapid disruption to labor markets (humanoid robotics and automation).
- On controversies: Gates expressed embarrassment over his past association with Jeffrey Epstein (saying meetings/dinners occurred but denying any involvement in Epstein’s criminal activities) and described continuing engagement with the Trump administration to advocate for global health funding (despite funding volatility).
Gates on AI and the Horizon 1000 pilot
- What it is: A pilot to deploy AI-enabled clinical assistants in 1,000 clinics, starting in Rwanda. Joint US$50m commitment from OpenAI and the Gates Foundation.
- Function: Voice and data interfaces to transcribe consultations, capture clinical data, run symptom evaluations, generate summaries for doctors, and support follow‑up (e.g., HIV patients, pregnant women).
- Language and data strategy:
- The foundation is funding data collection for many African dialects (~40 being benchmarked).
- Gates cited ~$4m per language to gather training data, which then can be used by major model providers.
- Connectivity:
- Two interaction modes: feature‑phone voice (possible but cumbersome) and smartphone/data connection (preferred for richer interface).
- Pilots will focus on clinics with data/smartphone capability rather than voice‑only setups.
- Expected benefits: make less‑trained practitioners far more effective, reduce paperwork, improve triage/decision quality, and extend continuous patient contact.
- Cautions: AI in clinics will focus on common, treatable conditions in these settings rather than rare diseases that require specialist care.
Aid, child mortality and philanthropy
- Historical gains: dramatic reduction in under‑5 mortality since 2000 largely due to vaccines, sanitation, and economic growth.
- Recent reversal: a small but worrying increase in child deaths in 2025 tied to disruptions in vaccine distribution, bed‑net distribution, and cuts to primary healthcare funding.
- Aid cuts: Gates points to significant cuts or uncertainty across multiple donor countries (including the U.S., U.K., France, Germany) and the operational impacts—hiring freezes, fewer deliveries, and program disruptions.
- Limits of private philanthropy: Even with large foundation budgets and Gates’s commitment to spend down over 20 years, private funds cannot substitute for the scale and steadiness of sovereign aid. Debt burdens in many African countries further complicate recovery.
Controversies and political engagement
- Epstein: Gates acknowledged past dinners and discussions with Jeffrey Epstein about fundraising but expressed embarrassment and emphasized there were no island visits and no women involved in his meetings with Epstein.
- Trump and U.S. administration:
- Gates described multiple meetings with President Trump (and other senior officials) focused on restoring or advocating for global health funding.
- He noted funding cuts (including to USAID and related programs) and some partial restorations; there remain discrepancies between congressional budgets and actual disbursements.
- Gates argues that engagement with political leaders is necessary to protect and advance global health programs, even when their behavior is unpredictable.
- On how CEOs should deal with Trump: Gates suggests engagement is necessary for policy work (particularly on global health), though he recognizes ethical and reputational trade‑offs for business leaders.
Notable quotes
- “The majority of Africans never meet a real doctor.” — highlighting the role of task‑shifting and the need for decision support.
- “What we’re talking about here is free intelligence.” — on AI’s potential to make expertise widely available.
- “I’m embarrassed that I spent any time with Epstein.” — on his past contact with Jeffrey Epstein.
- “People are afraid to speak out about being afraid to speak out.” — on the chilling effect of political climate and public discourse.
- “We’re spending 100% of my money… over the 20 years.” — on the Gates Foundation’s spend‑down commitment.
Implications / recommended actions (from interview analysis)
- For governments: stabilize and restore aid budgets and predictable funding streams (vaccines, bed nets, primary care) to avoid reversals in child‑mortality progress.
- For philanthropies: coordinate with governments and prioritize catalytic investments (R&D, data collection, pilot programs) rather than attempting to fully replace sovereign funding.
- For tech companies: invest in local data collection, language/dialect benchmarks, and infrastructure that enables safe, validated deployment in low‑resource clinical settings.
- For health implementers: prioritize clinics with reliable data connectivity for early AI deployments; focus on common, high‑impact conditions (HIV care, maternal health) and rigorous performance benchmarking.
- For policymakers and society: prepare for labor disruption from advanced AI/robotics, and urgently build safeguards against malicious AI use (especially in biosecurity).
This summary distills the interview’s factual points and the policy/technical implications Gates emphasized about AI in health, the limits of philanthropy, and the need for stable public funding to sustain progress in global health.