-
ChatGPT as good as doctors at suggesting diagnoses in ER
New research suggests that generative AI has the potential to speed up diagnosis and reduce waiting times in the emergency department.
-
Reward hacking: AI finds a way
Sometimes, the hacks AI comes up with are ingenious. Other times, they are a problem. But a lot of the time, they’re funny. Here are six that made me smile.
-
Virtual hugs work for real
Digital hugs can provide comfort and connection – but only if delivered with digital tact. Researchers present a framework.
-
You can’t fetch the coffee if you’re dead: an AI dilemma
An intelligent machine knows it can’t complete its objective if switched off. So, serving the coffee has a subgoal: avoid being switched off.
-
Why Facebook needs free speech
Without free expression, Facebook doesn’t have the diversity of opinions to fan the flames of engagement that drive advertising revenue.
-
Deepfakes trick your brain – just like text
New research shows that deepfakes trick your brain – but are no more misleading than simple text descriptions.
-
Data labellers: the invisible workers who make AI possible
Data labellers are the indispensable but unseen professionals who create AI training data – in sometimes murky conditions.
-
24. Our World in AI: Fun with fingers – Part 2
‘Our World in AI’ measures bias in Artificial Intelligence. Today: more fun with fingers as DALL-E grows up.
-
23. Our World in AI: School runs
‘Our World in AI’ measures bias in Artificial Intelligence. Today: school runs and a fair result.
Questions or suggestions?