Age of deepfakes means internet users must be more alert than ever, experts urge – World

The vertiginous evolution of artificial intelligence (AI) tools capable of generating convincing text, images and even live videos is allowing increasingly intelligent and specific scams, said experts in cybersecurity AFPurging Internet users to raise their guard.

In recent weeks, a “romance scam” of high profile in France, in which a woman forbid It is a goal for cyber attacks, “said Arnaud Lemaire of the F5 cybersecurity company.

One of Cyberattack’s best known forms is phishing, sending emails, text messages or other messages under false claims. Most try to make users take action how to click on a link, install a harmful program or disseminate confidential information.

Phishing and his Social Engineering cousin “Pretexting” together represented more than 20 percent of almost 10,000 data infractions worldwide, informed to the US telecommunications operator. UU. Verizon for the 2024 edition of their standard of their standard of The industry. Data violation research report.

The chatbots of Ia driven by large language models (LLM) save the attackers and allow more elaborate false messages, Lemaire said. They also mean that “if someone is writing a Phishing email … they can make the clues disappear completely”, which could give a non -native speaker of the destination language.

But text generators are only the tip of the iceberg AI.

For example, AI can “take advantage of all the data that have been violated in recent years to automate the creation of highly personalized scams,” said Steve Grobman, technical director of the MCAFEE security software manufacturer, to journalists last Thursday.

This is “something that only a few years ago would not be possible without an army of humans.”

‘Safe Word’

Instead of obtaining a quick score, the attackers often have the objective of gaining the confidence of the selected people in the target companies for months or years.

If an employee is successful, the attackers “could wait until this person becomes very influential or there is a good opportunity for them to extort the money” before exploiting the connection, Martin Kraemer said of the Knowbe4 cybersecurity training firm.

The bets were exhibited in February 2024, when the scammers cheated $ 26 million of a multinational company in Hong Kong. Police said a finance worker believed that he was in videoconference with the CEO and another company staff, everything, in fact, generated deepfakes.

“Deepfake’s latest video generation has reached a point where almost no consumer can notice the difference between an image generated by AI and a real image,” said Grobman.

Internet users must start applying the same skepticism to the video as many now make the fixed images, where “Photoshop” has become a verb, he added.

Before an alleged online news video, that could be as simple as verifying against a reliable source.

In personal communications, “I almost want to say that it is like BDSM, Bondage, where you have a safe word,” joked the f5 motto. “You tell yourself, here is the CEO that asks me to make a bank transfer of $ 25 million, I will bring something personal to make sure it is.”

Lemaire added that other tricks include asking a video caller to tell him his camera, something for now has difficulty recreating.

Cars horses

The online scam industry is so lucrative that “like other companies … there are supply chains and an ecosystem of tools to support it,” Grobman said.

Malicious rental programs include ransomware such as Lockbit, which can encrypt data on computers and threaten to release or eliminate it unless payment is made. One of his alleged developers was arrested in Israel in December pending extradition to America.

The AI ​​tools include one that allowed a McAfee researcher to replace his own face with the Hollywood star Tom Cruise in a video for only $ 5, Grobman said.

Even with all the new tools, “I am not too worried about the defense side that will overwhelm us,” said Kraemer de Knowbe4. It is “a tool that we can use for the attack and for defense,” he added.

However, the final defense line remains human for now.

“When we go from walking and riding cars, we needed to change the way we thought about transportation security … That is what consumers will need today, the same type of pivot,” Grobman said.



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *