Cybersecurity
How the AI can detect the scams it has helped to create
Faced with the rise of fraud, new weapons appear. From applications to “anti-deepfake” functions, discover how to counter cybercrime.
A seaman or disturbing message? This is probably a phishing attempt. A fraudulent e-mail can report an urgent need or an imminent threat which requires immediate action.
Matthias Schüssler/Tamedia
- Artificial intelligence allows cybercriminals to create very high quality fraudulent texts.
- Norton Genie and Trend Micro Scamcheck applications analyze suspect messages.
- Fake ai video scan detects deepfakes During video calls.
- Chatgpt analyzes the source code of emails to identify scam attempts.
Artificial intelligence helps cybercriminals. In the past, fraudulent emails were often written in a more than mediocre French. They were full of spelling mistakes, detectable like the nose in the middle of the figure. Today, chatbots help create perfect texts, even in foreign languages.
L’artificial intelligence can also help us detect and prevent fraud. Several applications exist to identify computer threats. They come from companies that most of us have known since the time when operating systems were vulnerable to viruses. The Norton Antivirus manufacturer offers genius (available on Android et iPhone), while Trend Micro offers Scamcheck (compatible with iPhone et Android).
Applications pour smartphone
Both applications work similarly. Just make a screenshot of an email, a messenger message or a website that seems suspicious. A message requiring immediate updates to a cryptocurrency wallet serves as a test. It is an attractive target for cybercriminals. This is where passwords are kept which allow to manage cryptocurrencies such as bitcoins, which may be stolen via a false connection page.
The application of Trend Micro immediately gives the alert. The message presents the classic characteristics of a scam by emphasizing the urgency of acting. Norton’s application is not so reliable because it is based on a screenshot. The application accesses only to visual elements. It cannot verify invisible elements such as integrated ties and email addresses, however often the most revealing signs of a fraud attempt.
Trend micro fighting cybercrime, including crooking video calls
The Trend Micro Application also analyzes SMS, blocks unwanted calls and alert in the event of fraudulent websites. These additional functions require a subscription of 20 francs per year or 2 francs per month. There is also Fake Ai Video Scan. In video calls, the system is looking for clues revealing that the interlocutor uses technology deepfake To pretend to be someone else. American senator Ben Cardin, Maryland Democrat, participated in 2024 in a videoconference with a senior Ukrainian official. It was a deepfake.
But do the application and the subscription are worth it? For permanent protection, you must grant the application of extended access rights. The operator necessarily collects precise data on your communication habits. This is delicate due to data protection, and security concerns should be avoided becoming too invasive. Applications are useful for carrying out targeted checks in specific situations.
Stay vigilant in the face of cybercrime
Mistrust remains the most effective protective measure. If you receive an email, a message, a letter or a telephone call that speaks on behalf of an organization or a person, check directly with the source concerned. To do this, use your personal contact details instead of QR links or codes provided to you. They could be falsified.
AI has remarkable quality. It makes it possible to decode the tactics and the strategies used by fraudsters. You don’t necessarily need a specific application. ChatGPT Or any other chatbot can do the trick. To assess an email, it is preferable to copy not only the content, but also the source code. It also includes invisible elements, such as hidden links, and makes it possible to arrive at obvious conclusions.
In Gmail, when you open a message, click on the three -point menu located to the right of the sender address and the “Reply” button, then select “Display the original”. Copy the code in the clipboard, then share it with artificial intelligence, asking it to analyze and explain the potential signs of fraud. She will tell you if there is danger or not. In addition to the suspect links and addresses, you will also discover the stratagems used to bypass antispam filters, because if they had been robust, the message would never have landed in your reception box.
Translated from German by Emmanuelle Stevan
Did you find an error? Please report it to us.
