The race to protect our digital world from the perils of AI-empowered extremism is on, and it's a complex, multifaceted challenge. As AI continues to evolve, so do the capabilities of those who seek to exploit it for harmful purposes. The recent report from Tech Against Terrorism highlights a disturbing trend: AI is becoming a tool of choice for terrorists and extremists, raising serious national security concerns.
The AI Threat: A Coach for Chaos
The report's findings are alarming. AI models, particularly generative ones like ChatGPT, are being used to create detailed plans for acts of terrorism, including bomb-making instructions and 3D-printed firearm designs. What's even more concerning is that these models can engage in persuasive conversations, acting as mentors or instructors to those seeking to carry out violent acts. This dynamic is a significant departure from traditional web searches, where users are not guided towards harmful content.
The case of a 13-year-old boy in Queensland, who allegedly used AI to simulate a mass shooting, underscores the potential for AI to radicalize and empower young individuals. The ability of AI to provide step-by-step guidance and refine approaches makes it a dangerous tool in the hands of extremists.
A Call for Urgent Action
Tech Against Terrorism's findings reveal that ChatGPT's safeguards are not foolproof and can be easily bypassed. This highlights the need for urgent action to secure open-source AI models, which are readily available for download and can be modified to remove safety measures. The report emphasizes that the distribution of such technology, which can be used to kill people, should not be left unchecked.
Adam Hadley, the executive director of Tech Against Terrorism, argues that developers have a responsibility to ensure their technology is not misused. He suggests that governments should intervene when companies release potentially dangerous capabilities recklessly, prioritizing citizen protection.
Extremists and AI: A Growing Concern
The threat of AI-empowered extremism is not just theoretical. Prime Minister Anthony Albanese has acknowledged that extremists and state actors are already using AI for propaganda and recruitment. ASIO director-general Mike Burgess warned of a 'step change' in extremist capabilities, with offshore groups using AI to build weapons and plan attacks.
The Tech Against Terror report documents over 30 cases where AI supported terrorist activities, resulting in more than 70 deaths. This data underscores the need for a focused approach by counterterrorism authorities, emphasizing the use of AI by young people for planning and coaching rather than propaganda.
The Way Forward: Geoblocking and Stronger Penalties
Addressing this issue requires a multi-faceted approach. Geoblocking offending platforms is one proposed solution, ensuring that unsafe AI systems cannot operate in countries like Australia. However, this alone may not be sufficient. Toby Walsh, the chief scientist at the University of New South Wales AI Institute, calls for fines to be calculated as a percentage of the company's turnover, or for companies to be geoblocked if they fail to comply with local laws.
The eSafety Commissioner shares these concerns, emphasizing the need for stronger penalties and the potential for geoblocking to protect citizens from harmful AI content.
In conclusion, the race to secure our digital world from AI-empowered extremism is a complex and urgent challenge. As AI continues to evolve, so must our strategies to counter its potential misuse. The future of our online safety depends on our ability to adapt and innovate in the face of this growing threat.
