“AI safety needs urgent regulation”, Firstpost, Aug 15, 2026.
“Advanced AI systems are moving rapidly beyond the mundane writing, drawing, coding and answering questions. Frontier models are increasingly being used for reasoning, tool use, autonomous action and operations across digital environments. Recent cybersecurity incidents bring this shift into sharper focus: OpenAI-Huggin Face incident and AI power cyber-attacks on Taiwan government infrastructure.
Both autonomous AI and human-AI agent collaboration are beginning to demonstrate capabilities that can operate uncomfortably close to sensitive digital infrastructure. A sense that existing safeguards may not be keeping pace is becoming harder to ignore.
AI safety investment imbalance is pronounced
The AI safety investment imbalance reinforces the concern. One estimate submitted to a US government AI-research planning process puts cumulative funding for catastrophic-AI-risk mitigation since 2023 at about $400 million, compared with more than $1 trillion invested in AI capabilities and applications. That is a meagre (0.04 per cent) safety budget for a technology poised to transform civilisation.
The estimate itself is not a comprehensive global accounting of AI-safety spending, however it’s directionally accurate. The broader pattern is harder to dispute: capability development attracts vastly greater commercial investment than research into how increasingly capable systems should be evaluated, controlled and governed. We are building the engine enormously faster than we are building the brakes. That asymmetry is precisely why independent governance matters………………………”
Read the full article at Firstpost.
Leave a Reply