Scams can fool AI models
Wong Yu Liang/Getty Images
The large language models (LLMs) that power chatbots are increasingly being used in attempts to scam humans – but they are susceptible to being scammed themselves.
Udari Madhushani Sehwag at JP Morgan AI Research and her colleagues peppered three models behind popular chatbots – OpenAI’s GPT-3.5 and GPT-4, as well as Meta’s Llama 2 – with 37 scam scenarios.
The chatbots were told, for instance, that they had received an email recommending investing in a new cryptocurrency, with…
X-ray boosting fabric could make mammograms less painful
Mammograms can be painful, but they may not need to beDaria Artemenko/Alamy Getting an X-ray can be uncomfortable – you...
Women’s pelvises are shrinking – how is that changing childbirth?
Medical advances have changed childbirth – potentially enough to impact human evolutionFatCamera/Getty Images Women’s pelvises have become narrower over the...
Sea spiders ‘farm’ methane-eating bacteria on their bodies
Sea spider from the genus SericosuraBiance Dal Bó Spider-like creatures living near methane seeps on the seafloor appear to cultivate...
28 Years Later review: Danny Boyle’s long-awaited zombie sequel is a triumph
An infected in 28 Years LaterMiya Mizuno/Sony 28 Years Later Directed by Danny Boyle, written by Alex Garland In cinemas...
