It sounds like the researchers at the Dawes Centre for Future Crime at University College London have done some pretty intense digging into the scary side of AI. You know, it’s kind of a double-edged sword, isn’t it? On one hand, AI’s doing all these amazing things for us, but then there’s this darker side where it can be twisted for some really bad stuff.

So, they’re talking about how AI could be used for crimes in different ways. It could be the actual tool for a crime, like when a hacker uses AI to break into systems or something. Or AI itself could be the target – like if bad guys try to mess with AI security systems. And then there’s using AI as a sort of backdrop for a scam, like making people think AI can do something it really can’t, and ripping them off that way.

The study pointed out 20 types of crimes that could be powered by AI, and they’ve sorted these into high, medium, or low concern categories. That’s pretty smart because it helps to know what we should be really worried about.

For the high-concern stuff, they’ve listed things like using AI to mimic someone’s voice or face – think deepfakes. Or, imagine driverless cars being used as weapons in terrorist attacks. Then there are tailored phishing scams, messing with important AI systems, doing blackmail on a massive scale, and spreading fake news with AI. That’s some heavy stuff.

The medium concern crimes are also pretty worrying. Like the misuse of military robots – that’s straight out of a sci-fi movie. And then there are scams pretending to be AI solutions, messing with AI learning data, cyber-attacks that are super targeted, using drones for crime, blocking people from online services, tricking face recognition, and playing the financial markets.

The low-concern ones are still not great but maybe less immediate. They’re talking about things like exploiting biases in AI algorithms, tiny robots for burglaries, slipping past AI security systems, faking online reviews, stalking with AI help, and creating fake art or music.

This whole thing is a wake-up call. While we’re all excited about the cool things AI can do, we’ve got to be on our toes about the not-so-cool things it can be used for. It means everyone, like the cops, the government, and even the folks making AI, need to keep an eye out and stay one step ahead of the bad guys.

Some key takeaways:

1. Dual Nature of AI: AI is a game-changer but has a dark side that can be exploited for criminal activities.
2. Versatile Criminal Tool: AI can be a tool for committing crimes, a target of criminal activities, or a deceptive context in scams.
3. High-Risk Crimes: Top concerns include AI-generated deepfakes for impersonation, using driverless vehicles in terror attacks, sophisticated phishing scams, disrupting critical AI systems, large-scale blackmail, and spreading fake news.
4. Medium and Low Threats: Medium concerns involve misuse of military AI, data manipulation, and AI in financial fraud, while lower risks include exploiting AI biases and creating fake digital content.
5. A Call for Vigilance: The study underscores the need for law enforcement, policymakers, and AI developers to be proactive in countering AI-related crime threats.

Always stay ahead a step by keeping an eye on such AI crimes by subscribing to our AI Crimes Newsletter. And keep yourself safe before falling prey.