Josephine Teo mentioned that stronger safeguards are being considered for high-risk AI uses as systems become more autonomous. This comes as AI systems grow in capability and independence.

There have been no reported cyber attacks involving unsupervised AI agents against government agencies. This is according to available information, which does not provide further details on the nature of the potential threats or the proposed safeguards.

Further information on the plans for stronger safeguards and the current state of AI security can be found in the original source.