Gates names four jobs safe from AI: useful as a compass, shaky as a promise of employment

By Emisoras Unidas · November 6, 2026.

Microsoft co-founder Bill Gates has named four areas where he believes people will remain indispensable even as artificial intelligence advances rapidly: programming, biology, energy and professional sports. According to Guatemalan outlet Emisoras Unidas, he made the remarks during an interview on The Tonight Show with Jimmy Fallon. He cited baseball as an example of an activity whose appeal is precisely its human element.

The piece sets out the reasoning attributed to Gates in each case. In programming, AI already writes code and spots bugs, but people are still needed to design systems, ensure their security and assess risks. In biology, AI speeds up the analysis of proteins, genetics and drugs, but it does not replace those who formulate hypotheses and validate results through experiments. In energy, it helps optimize grids and anticipate failures, but decisions on infrastructure and safety require specialist judgment. In sports, technology improves training and strategy, but it does not replace the value of watching people compete.

The article adds an important caveat. In a post published on August 26, 2026 on his blog Gates Notes, Gates warned of profound changes in the labor market over the next decade. He cited customer service, medicine, law, manufacturing and software development. He also acknowledged that some jobs would disappear and that new roles would not necessarily make up for all the losses. According to the outlet, his conclusion is that societies will have to decide which activities they want to keep under human responsibility.

Our take: the headline promises “four safe jobs,” but the material itself undermines that simple reading. Programming appears on the list of “safe” jobs and, at the same time, on the list of sectors set to be transformed. The apparent contradiction dissolves once you see that Gates is talking not about specific jobs but about functions: overseeing, validating, deciding. What survives is not “the programmer” as the role exists today, but judgment about what gets built and whether it is safe. For anyone writing routine code today, that is a warning, not a comfort.

The selection is not neutral either. Three of the four fields (software, biology, energy) coincide with areas where Gates has invested and voiced opinions for years. In general, lists of “AI-proof jobs” age badly: what looks like a safe haven today may stop being one within a few years. They are best treated as hypotheses, not guarantees.

The most valuable part is the underlying principle: deciding collectively what we want to remain human. That is where the most interesting question lies, and it is a matter of social choice, not just a technical one. Biology also ties into AI's great long-term promise: accelerating the fight against disease, with scientists directing and verifying rather than being replaced. Sports illustrates another idea: when abundance frees up time, what we value precisely because a person does it will gain in worth.

The short term, by contrast, will be less kind. If Gates admits that new jobs may not make up for those lost, the priority is not to hunt for a refuge occupation but to design the transition: training, protection and sharing of the gains. Without that, the promised abundance will arrive very unevenly.