Artificial Intelligence (AI) has many benefits. It can help us do our work – especially boring, repetitive tasks – more quickly and efficiently. But, it also helps scammers do their work better too. Here are five ways that hackers are using AI.
1. Bypassing CAPTCHAs
Those tests that websites use to verify that you’re human, whilst annoying, are used to protect customers from bots with stolen credentials. However, scammers are using machine learning (ML) and image recognition to overcome these, enabling them to do more credential stuffing attacks on a bigger scale.
2. Enhanced social engineering
The most successful fraud is one that’s believable. Scammers use AI to help them conduct more realistic phone scams. By using AI to scrape the internet for information about the target, they can tailor phone calls to make the scam seem more realistic.
3. AI-generated phishing
Similarly, scammers have also upgraded their phishing campaigns by using it to gather more information about targets so that emails can be personalised. It can also help scammers to scale-up these more effective campaigns quicker.
Deepfakes can be used as a blackmail tool, as scammers can manipulate videos to show real people doing things they haven’t. Random face generator websites can also provide scammers with faces to create false accounts. Some deepfakes are even sophisticated enough to trick image recognition software.
5. Voice cloning
This is particularly dangerous, as it enables scammers to replicate someone’s voice exactly. That means they can trick people into thinking their boss or finance department is asking them to transfer money to, or share confidential information with a scammer. AI enables this software to learn to converse in a more realistic way.
Whilst the use of AI makes it harder to detect scams, it’s important to still maintain common sense security practice. Always verify a caller or emailer independently, and if you believe you know the person, double check the number they’re calling from. If in doubt, contact the person separately on a number you know belongs to them.