Emphasis on ensuring effective safeguards and penalties against the misuse of AI
Concern has emerged that artificial intelligence could help terrorists make dangerous weapons.
According to recent research by Tech Against Terrorism, artificial intelligence could make information about developing biological weapons more accessible to terrorists and inexperienced extremists.
The report says that advanced open-weight AI models are particularly concerning because they can be downloaded and operated privately. Such models could provide dangerous assistance to individuals who previously lacked the scientific expertise required to carry out complex biological attacks.
Researchers say some AI models can be operated locally, meaning the information entered into them does not reach company servers. This could make it difficult for authorities to monitor their use. There are also concerns that the safety restrictions built into such models could be weakened or removed.
According to the research, in initial tests, around one-third of the responses provided meaningful assistance that could increase a malicious individual’s ability to prepare for terrorist activities.
The report calls for restricting access to sensitive information and biological materials, improving the detection of suspicious activities, and ensuring effective deterrents and penalties against those who attempt to misuse AI.
Also Read: Meta to Pay $18 billion over making children addicted to social media
