Artificial intelligence called a threat to national security: “Dive bomb on the crowd”

The creators of artificial intelligence need to abandon their “technological utopian” thinking amid fears that new technology could be used to exploit “vulnerable people” by terrorists, experts urge.

Jonathan Hall QC, whose role is to review the adequacy of anti-terrorism legislation, argues that the threat to national security from artificial intelligence is becoming more and more obvious, and the technology needs to be developed with terrorists in mind.

According to The Observer, Hall said that too much AI development is focusing on the potential benefits of the technology, while neglecting to consider how terrorists can use AI to carry out attacks.

“They need some horrible little 15-year-old neo-Nazi in the room with them to figure out what they could do. You need to protect yourself hard against what you know people will do with it,” Hall said.

An independent British government expert on counter-terrorism law has acknowledged that he is increasingly concerned about the ability of AI-powered chatbots to persuade vulnerable or neurodivergent people to carry out terrorist attacks.

“What worries me is the suggestibility of people when they are immersed in this world and the computer is turned off. The use of language in a national security context matters because, ultimately, language convinces people to do something,” the expert says.

Security services are believed to be particularly concerned about the ability of AI-powered chatbots to influence children, who are already a growing part of MI5’s list of terrorist suspects.

As calls for technology regulation grow following warnings from AI pioneers last week that it could threaten the survival of the human race, Prime Minister Rishi Sunak is expected to raise the issue when he travels to the US on Wednesday to meet with President Biden. and senior members of Congress.

Back in the UK, we will step up efforts to address the national security challenges posed by artificial intelligence in partnership between MI5 and the Alan Turing Institute, the national body for data science and artificial intelligence.

Alexandre Blanchard, Digital Ethics Fellow at the Institute’s Defense and Security Program, said his work with the security services shows that the UK is taking the security challenges posed by artificial intelligence extremely seriously.

“There is a greater willingness among defense and security policy makers to understand what is going on, how actors can use artificial intelligence, what the threats are,” says Blanchard. – Indeed, there is a sense of need to be aware of what is happening. Work is underway to understand what the risks are, what are the long-term risks and what are the risks for next generation technologies.”

Last week, Rishi Sunak said that Britain wants to become the global center for artificial intelligence and its regulation, insisting that this could bring “huge benefits to the economy and society.” Both Blanchard and Hall say the central issue is how humans retain “cognitive autonomy”—control—over AI and how that control is built into the technology.

The potential for rapid AI impact on vulnerable people alone in their bedrooms is becoming increasingly clear, Hall says.

Jonathan Hall notes that tech companies need to learn from the mistakes of past complacency – in the past, social media has been a key platform for sharing terrorist content.

Hall added that more transparency is also needed on the part of firms using AI technology, primarily about how many employees and moderators they have hired.

“We need absolute clarity on how many people are working on these things and moderating them,” he said. – How many people are actually involved when they say they have fences in place? Who checks the fences? If you have a company of two, how much time do they spend on public safety? Probably little or nothing.”

Hall said new laws may also be needed to combat the AI ​​terrorist threat to curb the growing danger of lethal autonomous weapons — devices that use AI to select their targets.

Hall said, “You’re talking about the type of terrorist who wants denial, who wants to be able to ‘fly away and forget.’ They can literally throw a drone in the air and fly away. No one knows what decision his artificial intelligence will make. For example, it could just be a dive bomb on a crowd. Do our criminal laws cover this kind of behavior? As a rule, terrorism is associated with intent; the intention of a man, not a machine.”

Deadly autonomous weapons – or “idling munitions” – have already been seen on the battlefields in Ukraine, raising moral questions about the implications of an autonomous air-killing vehicle.

“AI can learn and adapt by interacting with the environment and improving its behavior,” warns Alexandre Blanchard.