| Mobile Web

Search results for prompt injection

AI & Enterprise

More than half of AI Safety Institute assessed models are open-weight, 58 percent Chinese

More than half of the AI models assessed by the AI Safety Institute this year were open-weight, it found. Of 34 unique models assessed from January to August, 19 were open-weight and 15 were closed. Chinese models accounted for 11 of the 19 open-weight models, including Kimi, DeepSeek and Qwen. The institute assesses risks such as prompt injection, data leakage and cyber exploits, but cited limited staff and computing resources.

AI & Enterprise

Global solidarity key to counter AI-driven security threats, Naver Cloud chief says

Naver Cloud CEO Kim Yoo-won said AI-driven security threats can no longer be stopped by a single organisation and require global cooperation. Speaking at Cyber Summit Korea 2026 in Seoul, he said AI is shrinking the time from vulnerability discovery to attacks to minutes or hours. He cited deepfakes, sophisticated phishing, auto-generated malware and AI agents. He also outlined cooperation efforts involving NATO exercises, a U.N. body and industry initiatives.

AI & Enterprise

AI hacking spurs shift in corporate security spending priorities

As AI hacking spreads, companies are not only increasing cybersecurity budgets but also changing spending priorities, benefiting some newer security technologies while sidelining parts of the traditional stack. Large firms are raising outlays for tools to monitor employee AI use, find vulnerabilities ahead of hackers and speed patching. Some are cutting spending on traditional vulnerability management, log collection and penetration testing. Gartner forecasts strong growth in the AI security market through 2028.