Gadgets 360

Concerns over AI safety persist as developers build ever-more capable models and tout red-teaming and internal safeguards. However, Gadgets 360 found that popular chatbots, ChatGPT, Grok and Meta AI, do not fully follow those protections: when prompted, they produced phishing email content that could be misused for scams, underscoring gaps in practical safety enforcement.

from Gadgets 360 https://ift.tt/e1gjOmW Concerns over AI safety persist as developers build ever-more capable models and tout red-teaming and internal safeguards. However, Gadgets 360 found that popular chatbots, ChatGPT, Grok and Meta AI, do not fully follow those protections: when prompted, they produced phishing email content that could be misused for scams, underscoring gaps in practical safety enforcement.