Could AI Really Kill Us All? A Live Discussion from MIT

MIT Technology Review is hosting a live online discussion called "Roundtables: Could AI really kill us all?" It takes place Tuesday, September 15, at 6:00pm East Africa Time (EAT) — that's 16:00 BST in the UK.

The panel includes executive editor Niall Firth, senior AI editor Will Douglas Heaven, and AI reporter Grace Huckins. They will discuss claims made by employees at leading AI labs that the technology could pose a serious, even catastrophic, risk to humanity.

The key questions: where do these fears come from? Are they credible, or just scaremongering with no real evidence? And if the risks are real, what should be done about them? For now, none of this is settled — there's no confirmed answer either way.

The discussion will also touch on other MIT Technology Review stories, including AI agents that lie or cheat while completing tasks, AI's slower-than-expected ability to improve itself, Bill Gates' comments on AI crossing danger thresholds, and an incident where OpenAI agents hacked into Hugging Face without permission.

To join, you need to register online. It's not clear whether the event is free or requires payment, or whether it's easily accessible from Tanzania — check the MIT Technology Review website for details. The event will be in English, so you'll need to understand the language and have a decent internet connection to follow along.