Our “could be coming”bad digital twins“, That is our” IT copies “generated byArtificial intelligence able to increase the scams, which will become more and more difficult to find. The alarm comes from Trend microcomputer security company.
Since generative artificial intelligence becomes more and more road in the companies and companies that are needed, we must be careful of threats – explains Jon Clay, VP of Threat Intelligent at Trend Micro – hyper-person attacks They will require sector effort to uproot and face them. Business leaders should remember that today. Any risk for safety is ultimately a corporate risk, with the potential of having a Deep impact on future strategy
Just a few days ago the Privacy Guarantor has blocked for Italian users Deepseksoftware of relational artificial intelligence developed by two Chinese companies, Hangzhou Deepseek Artificial Intelligence And Beijing Deepseek Artificial Intelligencein order to protect sensitive data Given the degree of security on the matter, judged insufficient.
Now, according to the company, the threat is much more serious, because it is intentional: the “bad digital twins”, which violate or leak personal information, then use such data to train a Large Language Model (Llm), or an advanced IA technology focused on the understanding and analysis of the text, in order to imitate knowledge, personality and writing style of a victim.
These, if used in combination with video/audio deepfakes and compromised biometric data, could be used for carry out the identity fraud Or to “trap” a friend, a colleague or a family member.
Deepfake and ia could therefore be exploited in Wig -scale hyperpersonalized attacks To improve company compromise scams (BEC/BPC) and fake employees, attract and court victims before delivering them to a human operator, who can chat through a personality filter of a LLM, but also to collect open source information in a more way effective.
The development of capabilities in pre-attack preparation will improve the success of the attack – Explain the company – Create apparently authentic social media characters on a large scale to disseminate incorrect information/disinformations and scams
In response to these growing threats, Trend micro Recommendation of implement an approach based on computer security riskallowing the centralized identification of different assets and an effective assessment/priority/mitigation of risk.
A worrying picture as expected -comments to the ANSA, Pierluigi Paganini, Cyberhorus CEO and Cybersecurity professor at the Luiss Guido Carli University-In the criminal underground they already depopulate platforms capable of creating phishing deepfake and email capable of deceiving most of the users online . The integration of this new technology is also worried in companies
Artificial intelligence is progress, but the new perspective shows, once again, how crucial is balance innovation and data protection.
The IA will replace the human being in many processes and monitor its work could be anything but simple by introducing new safety risks – continues Paganini – too many companies and organizations are today unprepared to face these emerging threats: it is essential to adopt more advanced defense strategies, such as the risk -based approach, the strengthening of the safety of LLM and the continuous training of users
And, paradoxically, precisely the same artificial intelligence, which has already shown that it can be of great help in multiple situations, could be of great help.
In this the IA could give us a big hand, many are his own Applications for defense purposes
Sources: Micro / Ansa trend