Trends in social networks and The use of artificial intelligence has become attractive toolsbut also in potential threats to privacy. Transform a Studio Ghibli -style photo or rejuvenating the face may seem harmless, but these actions expose unique biometric data, such as face or fingerprints, which could be used by cybercriminals.
Cybersecurity experts compare the fact of scanning the face for an artificial intelligence video or taking viral photos with delivering the keys of their own home. Once these data are in the hands of third parties, they can be used to enter bank accounts, supplant identities or falsify documentation.
Mario Micucci, a cybersecurity specialist at ESET, warns: “Biometric identity is like the key to our house.” When uploading a photo to an unknown application or based on artificial intelligence, many people deliver their identity without knowing in which database is stored or what use will be given.
As an example, he mentions the case of Faceapp, popular during the pandemic. “We became younger or older versions without knowing that we were giving our face to developers. In addition, third -party images are created without consent, which violates privacy,” he says.
It also highlights the danger of Deepfakes, tools that manipulate voices and images to simulate false actions or statements. “These resources can be used for scamssupplant relatives or access bank accounts through biometric images, ”he explains.
Micucci alerts about cases such as Worldcoin, a company that offered cryptocurrencies in exchange for scanning the iris of users. “Many did not know what they would do with that information,” he says. This shows the importance of reading privacy terms before using any application.
“Our digital life revolves around information. Our face, footprints and personal data have value. That is why it is vital to think twice before sharing them,” he says.
The risk increases if the databases are violated. “If a hacker access them, you can use that information to enter systems, applications and bank accounts. Our face is an access key. Giving it without thinking is like delivering a copy of the keys to our house,” he warns.
The advance of AI, according to experts puts Guatemalan citizens at risk due to the lack of approval of a law that regulates its use and data obtained from the user. (Photo: Free Press: Shuttersock)
Real risks and lack of regulation of AI
Miguel Morales, director of the Digital Education Area at the Galileo University, also warns that tools such as Chatgpt 4.0, although useful to create photorealist images, can compromise biometric data, such as the face, which are used to unlock devices and access confidential information.
“The problem is aggravated with free versions, which do not guarantee data protection. Identification documents, financial or confidential information,” Morales explains.
It ensures that a single image is not enough to collect all biometric features, but that a constant interaction, such as a direct video or scan, or even the sale of data, can be sufficient to capture them.
One of the biggest risks, he warns, is the cloning of voice. “With only five seconds of audio, it is possible to replicate a human voice and use it for scams through false calls.” In addition, there are already videos generated with cloned image and voice, used to deceive.
In 2024, CNN reported a band dedicated to romantic scams through Deepfakes. They simulated video calls with fictitious women who convinced the victims to invest in cryptocurrency plans. “Those responsible used videos, audios and other files manipulated to pretend sentimental relationships,” he said Huffpost.
In Guatemala, Morales points out that there is no specific legal framework on artificial intelligence. The 6347 initiative, still pending in Congress, contemplates regulatory aspects related to AI and National Security.
Other emerging risks are injection attacks. This technique allows attackers to supplant identities against facial verification systems with false faces generated in real time. Thus, they cheat cameras and programs that validate real identities.
There are also attacks with identity templates, where documents are falsified through AI, using data extracted from social networks or databases to supplant identities for criminal purposes.
Tools such as Chatgpt 4.0, although useful for creating photorealistic images, can compromise biometric data. (Photo: Free Press: Shuttersock)
Recommendations to protect yourself
From ESTT Latin America they recommend the following measures to reduce risks when interacting with artificial intelligence applications:
- Review privacy policies: Know how information is handled before uploading personal data
- Verify legal support: Make sure the platform complies with laws and ethical principles for data protection
- Use official applications: Avoid unsecured versions that could contain malicious software
- Up only safe images: prefer illustrations or landscapes without copyright, instead of personal images