Artificial intelligence, an Italian start-up against hackers stealing company data
The system developed is called IdentifAi, the antagonist of Ai: it unmasks those who enter online meetings in disguise by tracing emails, phone numbers and confidential information
by Sara Monaci
A start-up created by a pool of young engineers has emerged in Italy that uses artificial intelligence to stop industrial espionage. But not the traditional kind of industrial espionage: we are talking about those frauds that exploit video-conference meetings to commit fraud and steal company data.
The start-up is called IdentifAi and is marketing a system that is in fact the antagonist of Ai. If artificial intelligence can in fact become a tool in the hands of a hacker in the service of corporate fraud, the system set up by IdentifAi unmasks it. This innovation is the first in Italy - there are few similar realities in Europe - so much so that the start-up has been selected among the 100 best Ai companies in the world to participate in the Google Gemini Founders Program, the programme for artificial intelligence start-ups.
Ai tools now make it possible to reproduce the voice and likeness of a participant in video conferences, making fraud, espionage and sabotage possible. This type of fraud is not yet widespread in Europe, it is much more so in the United States and Latin America, but it is spreading everywhere. What happens is this: a hacker enters e-teams in disguise and then asks for data to be sent, and once he has obtained the e-mail, he is able to trace it back to the databases. Or he can simply get phone numbers of one of the participants, who can be contacted using a false identity to request sensitive information.
It may sound like a game, but it is a real danger capable of compromising business functions, interfering with the activities of institutions and even undermining national security. So many companies have been affected in the last 12 months. According to a late-September analysis by Gartner, a multinational IT strategy consultancy, 29 per cent of corporate cybersecurity managers said their organisations had experienced attacks on their generative AI applications in the space of a year, while as many as 62 per cent of companies had experienced deepfake attacks conducted through social engineering or automated exploitation processes.
IdentifAi, founded by Marco Ramilli and Marco Castaldo, has therefore created a 'degenerative' artificial intelligence, the IdentifAI Agent, capable of verifying videoconferencing environments - Zoom, Google Meet and Microsoft Team - by exploiting machine learning models to analyse video streams and detect digital manipulations in real time.
