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…
How government use of AI could hurt democracy
AI could automate some government paperwork, but it comes with serious risksBrett Hondow / Alamy Many countries are exploring how...
Surgical robots take step towards fully autonomous operations
A surgical robot operating on a dead pigJuo-Tung Chen/Johns Hopkins University An AI-powered robot was able to remove a gall...
Fig trees may benefit climate by turning carbon dioxide into stone
Fig trees may be especially good at removing carbon dioxide from the atmosphereRaimund Linke/mauritius images GmbH/Alamy Some fig trees can...
3D printing could enable a safer long-term therapy for type 1 diabetes
People with type 1 diabetes can’t produce enough insulin to regulate their blood sugarHalfpoint Images/Getty Images Researchers have 3D printed...
