The concept of artificial intelligence (AI) is becoming increasingly prevalent, and it seems like there is hardly anyone who hasn't used ChatGPT or watched a deepfake video. But like any good thing, the technology designed to assist humans is sometimes exploited for other purposes.
More stories:
Artificial intelligence is a learning technological system that mimics human intelligence and interaction, thought patterns and human-like speech. It performs tasks through self-learning from its own mistakes, improving and upgrading based on the information it gathers, all without the limitations of the human brain.
Those who have already begun to adopt artificial intelligence and use it for negative purposes are criminal organizations that delve deeper into its revolutionary technology. Recently, law enforcement has started studying the tool, although they are currently refraining from developing it due to a recent scandal in which the police were accused of using cutting-edge spyware to hack civilians’ phones without a court order and the subsequent moratorium on police use of such technological tools, including AI.
"It's a phenomenon that is gaining momentum among criminals and serves as fertile ground for extortion with threats to civilians," says Chief Superintendent David Katz, commander of the Cyber Unit in Lahav 433, popularly dubbed the “Israeli FBI”, in an interview with Ynet’s sister publication Yedioth Ahronoth. "Criminals are utilizing it, primarily in the realm of fraud. It's a familiar threat to us."
"Criminals impersonate CEOs of various companies. Today, you can record a voice message with your own voice and ask the artificial intelligence to change it completely to match the voice of a CEO of a large company. This way, the criminal can contact the company's CFO, pretend to be the CEO, and execute financial transfers from the company. This primarily happens through Israeli criminals targeting international business companies."
Artificial intelligence has the ability to provide authenticity to impersonators. "It's easiest to impersonate familiar individuals, whose voices can be sampled from past interviews in the media, and then 'dress' them onto the voice of someone else that the impersonator wants to mimic," says Katz.
"Artificial intelligence can gather information about a person from the Internet, learn it and even anticipate the responses that person would give in various situations. This enhances the ability to carry out perfect impersonations."
Currently, law enforcement primarily uses artificial intelligence for data learning and organization. "When you want to learn about processes taking place in the world, it can be very helpful. For example, if we want to have a broad overview of criminal phenomena in various countries worldwide, artificial intelligence, using ChatGPT, can provide that overview within seconds.
If I were an analyst who had to gather that information myself, it would take at least a week. It streamlines the process, but we are also very cautious about the information provided by artificial intelligence and make sure to cross-reference it because there is a lot of made-up data. Therefore, it remains an assisting tool."
According to Katz, soon anyone with some understanding of technology would be able to become a hacker. "You don't need to sit and write code or be a highly skilled programmer to carry out a cyberattack. Artificial intelligence can do it for you. And now we're seeing a new dimension of criminals who, until recently, lacked the ability to operate in this field, and suddenly it's accessible to them.
We'll have a world of hackers who are not truly hackers but criminals. This allows them to infiltrate computers, steal material from compromised systems and demand ransom payments for it. They can breach mobile phones and eavesdrop on conversations."
Katz also discusses the challenges in enforcing technological offenses as artificial intelligence develops and its uses expand. "The ability of law enforcement to enforce lies in the traces that criminals leave behind, which can later be translated into evidence against them.
However, artificial intelligence can completely erase the traces of criminals. All one needs to do is write a prompt. It is a very significant threat that will prevent the building of visual evidence infrastructure against criminals who commit offenses using artificial intelligence."
Another threat, according to him, lies in the fact that "artificial intelligence is merely a good tool intended to assist humanity, but the problem lies in the fact that it relies on publicly available sources of information. We already see high-level criminals, some of them state-sponsored hackers from enemy countries like Iran, entering publicly available databases and altering lines of code there. It's not science fiction; it's already happening.
Therefore, as law enforcement, we are very cautious since we also began using artificial intelligence. We don't want artificial intelligence to fully understand the needs and interests of the Israeli police, as it could potentially fall into malicious hands."