The Bill Gates human reserved concept, outlined in a 6,000-word essay published on Gates Notes on 26 August 2026, goes further than most observers had anticipated: Gates argues that society should actively protect up to 40% of current jobs from AI displacement, and pair that protection with a new tax on the AI outputs that replace human labour.
It is a bolder policy position than his previous writing suggested, and one that will force a sharper debate than the usual hand-wringing about automation.
What ‘human reserved’ actually means
The phrase, Gates writes, is borrowed deliberately from the idea of nature reserves. ‘We could put buildings and roads, but we choose not to because the loss would be too great.’ His examples lean heavily on care: the nurses and doctors who sat with his father through Alzheimer’s disease, the clinician who delivers a terminal diagnosis in person. ‘Imagine a robot giving you the awful news that you have an incurable disease. There’s no technical reason why it couldn’t. Yet it shouldn’t.’
The argument is not purely sentimental. Gates proposes taxing revenue from companies that use AI outputs to displace workers, with the proceeds funding a broader social safety net. According to Reuters, he specifically envisioned levying charges on AI ‘tokens’, the computational units behind model outputs, as a mechanism to redistribute gains from automation. Whether that is administratively feasible is a question the essay does not fully resolve, but the direction of travel is clear: Gates thinks capital that displaces labour should pay for the consequences.
His foundation is putting money behind the optimistic version of that bet. The Gates Foundation is aiming to spend $200 billion by 2045, with an increasing focus on AI. It is already working with Anthropic on AI-powered health checks for pregnancy and with OpenAI to support healthcare workers in Rwanda.
Gates on governance: the harder problem
The essay’s governance section is, if anything, more arresting than its labour market prescriptions. Gates compares the institutional challenge to the post-9/11 reorganisation of the US government and concludes that AI ‘will require much, much more.’ The list he produces is not rhetorical: employment, education, taxation, energy, elections, water, public health, the financial system, law enforcement, transport and IT infrastructure. ‘It is fair to wonder whether the world’s institutions are up to the task.’
He also warns of dangers that extend well beyond the economic. The Independent reports that the essay flags AI’s capacity to enable biological attacks, including the synthesis of dangerous molecules. That risk sits alongside a binary framing Gates deploys for equity: ‘AI will either be the greatest equalizer ever invented, or the worst source of injustice.’
On international governance, Gates is direct. His staff are arranging a meeting with President Xi Jinping in November. According to the South China Morning Post, Gates wrote that countries hosting leading AI developers and controlling critical parts of the supply chain ‘should begin meeting now to set up shared norms, before competitive pressure makes it harder for them to cooperate.’ The last time Gates met Xi was three years ago, when he became the first foreign entrepreneur the Chinese president had received since the Covid pandemic.
My read is that the Xi meeting is the most consequential item in the essay, even if it gets less attention than the human reserved framing. Bilateral AI norm-setting between the US and China is where the rubber meets the road on every risk Gates identifies, from biosecurity to labour displacement to election integrity.
The Epstein context
The essay arrives weeks after the Gates Foundation published the findings of an external review into its historical ties to Jeffrey Epstein. The review, conducted by law firm WilmerHale over approximately five months and involving more than 50 interviews with current and former foundation personnel (including Gates himself), found no evidence that Epstein was paid by the foundation and no evidence the foundation was contemporaneously aware of Epstein’s use of its name to advance his conduct.
According to AP News, foundation conversations with Epstein focused on a public health fund that never materialised and a grant to the International Peace Institute. The WilmerHale review, whose findings were announced on 21 July 2026, documented roughly 30 meetings between Epstein and foundation leaders and staff between 2011 and 2014. The governing board has since approved a centralised vetting process for brokers and donor advisers, with added scrutiny for meetings proposed at private residences.
Gates, who has not been accused of involvement in Epstein’s crimes, remains chair of the $89bn (£65bn) foundation.
The essay is his most substantial public statement in three years. Whether it shifts the policy conversation, or gets absorbed into the noise of an AI debate that moves faster than any governance framework can follow, may depend on whether that Xi meeting actually happens in November.
