Over the weekend, several of the most prominent people in artificial intelligence publicly backed calls to slow the technology down. Anthropic CEO Dario Amodei raised the concern, and OpenAI's Sam Altman and Elon Musk both said they agreed.

It follows an open letter, organised by the Future of Life Institute, calling for a prohibition on developing superintelligence until it can be done safely, now signed by more than 800 people, including Nobel laureates, former military chiefs and public figures.

President Donald Trump dismissed the calls on Sunday. If this has been filling your feed and unsettling you, here is what is actually happening.

Three things are worth holding onto. First, this is a disagreement, not a consensus. The people who build these systems do not agree on how dangerous they are, which is precisely why it is being argued about in public rather than announced.

Altman told Fortune that OpenAI has discussed pausing tests to make the technology safer, while Trump's science adviser David Sacks replied, in effect, that if the labs think they should slow down they should just do it and stop asking permission.

We’ve been here before. In early 2023 the same institute and many of the same signatories called for a six-month pause on models more powerful than GPT-4. It was ignored, development continued, and the specific catastrophes people feared did not arrive. That is not proof they never will. It is a reason not to treat this weekend as a countdown that has just started.

The fear feels new because it is suddenly everywhere at once, arriving on top of the Middle East conflict, high prices, worry about data centres draining power and water, the China rivalry and headlines about robots. The AI story is not separate from that pile. It is being read through it.

NPR reported Trump's response, that dire warnings are exaggerated and that "negative forces" are raising things that will not happen, alongside Altman and Musk backing Amodei's concerns. Bloomberg carried Trump's framing that whoever wins AI wins and that he wants the US to keep its lead over China. NBC News set out the open letter itself, which asks for a halt only until there is both scientific consensus that superintelligence can be built safely and public buy-in.

Almost nobody reading this is training a frontier AI model, and that is the useful vantage point. The entire fight is between a handful of US labs and the US government over how fast to move and who gets to set the rules. South Africa is not at that table, which means whatever pace, safety standard or guardrail emerges, we will inherit it rather than shape it, the same position we are in on age verification and teen safety features.

So the calm reading is this: the tools you use today, the assistant that drafts your emails or checks your code, are not what anyone is warning about. The warning is about future systems that do not exist yet, and serious people genuinely disagree about whether they will be dangerous. Be wary of anyone, in either direction, who tells you they know for certain. Nobody does, and confident certainty is the one thing the actual experts are not offering.

For an operator, the practical move is unchanged: use what works, keep a human checking the output, and do not bet your business on a specific prediction about a technology its own builders cannot agree on.

You might also like our piece on ChatGPT for Teens and who sets the rules for AI we use here, the GPU build in Centurion and the power question behind data centres, and how SA enterprise AI adoption keeps outrunning its own strategy.

Get more SA tech and business news and subscribe to The Open Letter.