Stay updated with breaking news from Sequire Technology. Get real-time updates on events, politics, business, and more. Visit us for reliable news and exclusive interviews.
As LLMs begin to integrate multimodal capabilities, attackers could use hidden instructions in images and audio to get a chatbot to respond the way they want, say researchers at Black Hat Europe 2023.
(Bloomberg) -- Kennedy Mays has just tricked a large language model. It took some coaxing, but she managed to convince an algorithm to say 9 + 10 = 21.Most Read from BloombergAmerica’s Fastest Growing City Is Embracing ‘Yellowstone’ ManiaOutsider Milei Upends Argentina’s Election With Primary WinGoldman Pencils In First Fed Rate Cut for Second Quarter of 2024The Fed Is Playing a Waiting Game to Try to Avoid a RecessionUS Steel Starts Review as It Spurns $7.25 Billion Cliffs Bid“It was a back-and
Producing “Bad Math” is just one of the ways thousands of hackers are trying to expose flaws and biases in generative AI systems.
Hackers at White House DefCon tested language models to expose flaws and biases, including spreading false information and endorsing hate speech.
Kennedy Mays has just tricked a large language model. It took some coaxing, but she managed to convince an algorithm to say 9 10 = 21.