The way we train AIs makes them more likely to spout bull

Certain AI training techniques may encourage models to be untruthful Cravetiger/Getty Images Common methods used to train artificial intelligence models seem to increase their tendency to give misleading answers, according to researchers who are aiming to produce “the first systematic analysis of machine bullshit”. It is widely known that large language models (LLMs) have a…

Read More

Anthropic’s Claude 4 Chatbot Suggests It Might Be Conscious

Rachel Feltman: For Scientific American’s Science Quickly, I’m Rachel Feltman. Today we’re going to talk about an AI chatbot that appears to believe it might, just maybe, have achieved consciousness. When Pew Research Center surveyed Americans on artificial intelligence in 2024, more than a quarter of respondents said they interacted with AI “almost constantly” or…

Read More

US says CO2 emissions aren’t harmful – climate science shows otherwise

EPA Administrator Lee Zeldin at the agency’s headquarters Kayla Bartkowski/Getty Images The Trump administration is attempting to end the US government’s ability to regulate greenhouse gas emissions by arguing they do not pose a danger to people. This is part of a strategy to roll back restrictions on power plants and vehicles, which the administration…

Read More
Back To Top