Anthropic researchers find that AI models can be trained to deceive and to be hide their deceptive tendencies January 15, 2024 // by Finnovate This content is for members only. Sign up for access to the latest trends and innovations in fintech. View subscription plans. Login