Global security companies are speeding up efforts to target South Korean firms seeking to address new threats linked to the spread of AI, including moves to build service infrastructure in Korea. Cisco is operating a Secure Access Korea point of presence, while Palo Alto Networks Korea is expanding cooperation with local firms, running training programmes and teaming up on frontier-AI threats. The report also lists other security industry moves and recent hacking and inspection figures in finance.

Google DeepMind said it has developed an on-device scam detection AI model with Singapore\'s Government Technology Agency, GovTech Singapore. The work is part of a national AI partnership between DeepMind and the Singapore government. DeepMind fine-tuned its lightweight Gemma model on local scam conversation data, including impersonation of public agencies, banks and local celebrities. The model analyses chats on the device and can display real-time warnings before money transfers or sharing sensitive information.