When Language Learns Mistakes
Illustration for the article "How business can keep natural language processing from going bad", published in Secure Futures, Kaspersky’s magazine on technology and digital security.
The article looks at the dark side of large language models – from biases that reinforce stereotypes and discrimination to the risk of exposing sensitive data absorbed during training – and outlines strategies businesses can use to make their use of AI more transparent and responsible.
Work