Forum: Guarding against dark side of AI requires comprehensive safeguards
Sign up now: Get ST's newsletters delivered to your inbox
Artificial intelligence (AI) has brought enormous benefits, from improving productivity to sparking creativity and driving innovation across many sectors. Yet, as with any powerful technology, there is a darker side that demands our attention.
A worrying scenario is when malicious actors deliberately build “poisoned” AI systems.
On the surface, these tools can appear helpful – while subtly weaving toxic narratives, misinformation or extremist ideas into everyday interactions. Over time, such influence can reshape beliefs without users even realising it.
This is not far-fetched. Studies show mainstream AI systems sometimes spread false information unintentionally, while cyber criminals have begun exploiting AI for scams and disinformation campaigns.
Globally, policymakers are responding and Singapore has taken a proactive, multi-pronged approach. National efforts now include the Global AI Assurance pilot, safety testing guidelines for generative AI, expanded skills training and regular industry consultations to keep pace with fast-evolving threats.
Research shows extremist groups are already exploring AI, while AI Singapore and the Infocomm Media Development Authority have launched new educational campaigns so all citizens can use, question and safeguard AI tools responsibly.
As AI grows more sophisticated, safeguards must go beyond technology to strengthen digital literacy and critical thinking at every level. Enforcement against malicious use is vital, but so too is supporting resilience in individuals, organisations and society.
AI is a powerful tool for good. But if weaponised, its ability to quietly influence and “poison minds” at scale may prove one of the greatest challenges yet for society.
Adren Sim
