In a new essay published Tuesday, Bill Gates argues that society needs to start making some difficult decisions about artificial intelligence now, rather than waiting until the technology is capable of doing substantially more of our work.
One of his most interesting suggestions is something he calls “Human Reserved.” The idea is essentially that certain jobs or activities would remain in human hands, even if AI eventually becomes perfectly capable of doing them.
“We should designate some jobs as Human Reserved,” Gates writes.
It’s an interesting concept considering much of the conversation about AI and jobs has focused onwhat AI can replace. Gates is asking a different question: What shouldn’t we allow it to replace and what should we maintain as human?
Some jobs may be worth keeping human
In his essay, Gates compares the idea to setting aside land as a nature reserve. We don’t preserve a forest because humans are incapable of building on it. We preserve it because we’ve decided there’s value in leaving it alone.
He argues we may eventually need to make similar decisions about work.
Child care is one example Gates has raised. Even if AI and robotics eventually reach a point where machines can safely look after children, parents may still decide that caring for kids is something humans should do.
Jury service is another. Those examples make “Human Reserved” sound relatively straightforward, but the idea gets considerably more complicated when you start applying it to professions that AI is already entering.
Education and health care, for example, could use AI while still keeping people at the center. A teacher might use AI to create personalized lessons or identify where a student is struggling without handing the entire classroom over to an AI tutor. A doctor could use an AI system to analyze medical information while remaining responsible for the patient’s care.
In other words, reserving something for humans doesn’t necessarily mean banning AI from it. It could mean deciding where AI remains a tool rather than becoming the worker.
Gates doesn’t think retraining everyone will be enough
Gates’ argument goes beyond the prediction that AI will eliminate some jobs while creating others.
For years, one of the most reassuring responses to fears about automation has been that workers can simply retrain for whatever jobs come next. Gates isn’t convinced that will be enough this time.
If AI systems and robots eventually become capable of performing an enormous range of physical and cognitive work, there may not always be another economically necessary human job waiting on the other side of retraining.
That’s a much bigger problem than learning how to use ChatGPT to do your job better. As Gates argues, governments may eventually need to intervene in other ways, including potentially taxing AI tokens or robots. The idea would be to change some of the economic incentives around replacing workers while generating money that could help people affected by automation.
Those proposals are a long way from becoming policy, and Gates isn’t presenting “Human Reserved” as a fully formed regulatory framework.
The takeaway
AI capability shouldn’t be the only test of where the future of work is headed. We’ve become accustomed to judging AI by capability such as can it write this? orcan it diagnose that? And while every improvement gives us another demonstration of something AI can now do, the capability doesn’t automatically answer whether we want AI doing it.
That’s what makes Gates’ “Human Reserved” idea interesting to me. It introduces a second test. We need to start asking do we want AI to do this for us? Frankly, believe those two questions have very different answers.
- Perplexity’s new local-first AI runs on your PC — and asks before using the cloud
- ChatGPT can learn your repetitive tasks by watching you — here are 7 prompts to try
- The best way to use ChatGPT for writing isn’t writing — 7 prompts I use to make it my creative partner
