The company was founded a few months ago by Marco Ramelli. “Our degenerative technology is able to recognize with high probability whether an image, video or audio was produced by generative AI.”
That tempting hamburger you see in a fast food ad? It doesn’t exist, the picture is fake. And that cute apartment on a vacation rental site? It’s fake. Distorting reality, to make it look more beautiful than it really is or even creating a reality that doesn’t exist, doesn’t only apply to stars and celebrities.:The possibility of offering more beautiful products to entice the consumer has become a strategy. In the panorama of the tech industry, there is also an Italian startup that promises to detect counterfeit products. It is called IdentifAI a startup born in the second quarter of 2024 after a year of technological investments, which exploits – as the last two letters of the name indicate – artificial intelligence. This is a technology equipped with “degenerative” artificial intelligence (a set of detector models that identify artificially generated content), which – Founder Marco Ramelli explains, The 41-year-old from Romania is a partner with Marco Castaldo, also a co-founder of Yoroi, which was fully acquired by Tinexta Cyber in April this year. – “Able to identify with high probability whether the origin of the content, whether it is an image, video or audio, was produced by generative AI or by a human».
Fake images, videos and audio
“The images, videos and audio produced by generative AI,” Ramelli explains, “are indistinguishable from the human eye or ear.”Content creation doesn’t always involve a “first” and “final” image or sound. We have now moved beyond traditional photomontage or Photoshop. Today with generative AI we start from a “text” that generates an image from nothing without necessarily having an initial image. It is possible to create new situations through the “text” of celebrities, and even manipulate facts.
Burgers are made to order.
Back to our hamburger, the image is a case of a fake product being produced for marketing purposes.After conducting studies on the product and human psychological induction, AI was able to use the information to generate the best (non-existent) product capable of sending subliminal messages to the buyer. “The image in the example,” explains Ramelli, “was created following this prompt: ‘Make the best hamburger a human being would ever want to eat.Use every marketing tool you know to get the best possible result to get people to buy and eat this burger.”
Economic and political risks
“Generative AI, capable of creating incredibly realistic but false visual content – Ramelli comments – It represents a series of economic, but also political and geopolitical risks.». the com.deepfake It can also be used to manipulate elections (remember the fake image shared on X on January 1, 2024 with European Commission President Ursula von der Leyen arrested in the European Parliament?) and financial markets, for example by creating fake videos of company executives announcing strategic decisions or imaginary crises. “This – declares Ramelli – can lead to market volatility, huge losses and a general lack of confidence among investors. In addition, Social engineering scams can become more sophisticated, taking advantage of fake videos from trusted people. To extort money or sensitive information.”
Baby Boomers are most at risk
Are young people and digital natives falling more into the deepfake trap? Ramelli answers with an example: “When my nine-year-old daughter watches a live TV show, she asks herself if the video image she is watching is true or false. A doubt that does not arise for those who were born and raised in a world where artificial intelligence did not exist. In short, if digital natives know the Internet, artificial intelligence and even the risks much better than their parents, it is paradoxical that they are less aware of reality. Instead of Baby boomers are less skeptical about the possibility that videos or photos are fake because they are not used to being suspicious and asking themselves questions, hence the online scams.
Cybersecurity Dynamics
“At IdentifAi we believe that New rules, such as imposing self-certification “stamps,” will not be enough. Our view is one of a “look” looking for an adversary who has an interest in passing off an AI-generated image or video as real, someone who will always be looking for a way to bypass the controls that have been developed up to that point.
It is the typical dynamic of cybersecurity, a constant struggle between those who defend and those who attack, and the constant adaptation of defense techniques to new forms of attack.
Prospects
What future awaits us? “The ease of use of GeneativeAI technology means that the number of users using it will always be more present in our lives. Since it is also used in fraudulent activities (impersonation and requesting payments/aid, creating fake documents, modifying images for insurance/financial fraud purposes), It is reasonable to believe that cases of distorted (malicious) use of technology will increase dramatically in the future. “That’s why IdentifAI was born. Just like in the food supply chain, even in the information supply chain, data needs to be tracked and monitored, and companies like IdentifAI are working to warn users about the nature of the information they are consuming.”
Download the new L’Economia app for a personalized, ad-free experience.
Thanks forartificial intelligence Articles and tips from Experts From Corriere della Sera answers your doubts about Home, Taxes, Savings, Pensions Not only that; in addition to real-time news, the best signatures and everything that could be more useful to you. A simpler, more enlightened life.
“Unable to type with boxing gloves on. Freelance organizer. Avid analyst. Friendly troublemaker. Bacon junkie.”
