The second category is deepfake video scams or fake ads. In 2024–2025, campaigns have exploded where supposed celebrities, politicians or business leaders recommend “miracle” investments, cryptocurrencies or financial products. Investors and consumers see videos with real faces saying things they never said. Recently, waves of deepfakes of public figures have been reported promoting investment frauds, precisely because people trust the authority of the person. Here the trap is psychological: if so-and-so says it, it must be serious. And it isn’t.

Third category: hyper-realistic, multilingual phishing. In the past, fraudulent emails were obvious because of spelling mistakes. Now, with generative AI, scammers write like a real company, in your language, with flawless style and personalized references. Europol and cybersecurity firms note that AI can generate perfect messages in seconds, adapted to each victim’s profile. A typical example: you receive an email supposedly from your bank or a major shopping platform saying there is suspicious activity, and you must “verify your identity”. The link takes you to a cloned website identical to the real one. AI has produced the site, the texts and the support chat. You enter your data. Account taken over.

Fourth category: cloned online stores and fake shopping ads. E-commerce fraud has multiplied because AI lets criminals create, in minutes, pages almost identical to well-known brands, with realistic photos, fake reviews and targeted advertising. In recent campaigns (especially during shopping seasons), waves of fake websites have been detected driven by deepfake ads or AI-generated messages. Example: you see on Instagram an offer from a familiar store with “50% off only today”. The ad looks real. The website is identical. You pay. Nothing arrives and your data is exposed.

Fifth category: sextortion and manipulation with fake images. AI allows criminals to create fake intimate photos or videos (or highly convincing edits) to blackmail victims. Although this threat is less “massive” than phishing, its emotional impact is enormous. The victim receives a message: “we have your images, pay or we will send them to your contacts.” Many times,the images are not real, but fear paralyzes. Regulators such as Ofcom in the UK have highlighted the growing presence of harmful deepfakes and related scams, with a high level of realism perceived by users.

The most dangerous thing about all these scams is that they have no “victim profile”. Older and younger people fall for them, executives and students alike. Because the decisive factor is not the victim’s intelligence but their emotional state and the urgency of the moment. AI is designed for that: to sound and look authentic precisely when you cannot verify.

What can you do to avoid falling for it? The golden rule is simple: distrust urgency. Every AI scam pushes you to act fast. If someone asks you for money “right now”, break the script. Breathe and verify through another channel. If it is supposedly your child, call their real number yourself. If it is your boss, write through your usual internal channel. If it is the bank, go into the official app, never through links. And if the message comes with video or voice, remember that today that no longer proves anything.

And if you have already fallen for it, what matters is to act calmly and quickly. First, stop the damage: block cards, change passwords, activate two-factor authentication, notify your bank or platform to freeze movements. Second, collect evidence: screenshots of emails, chats, numbers, links, audios, ads, transfer receipts. Third, report it: police or specialized prosecutors, and also the platforms where it happened WhatsApp, Instagram, TikTok, email provider, because their reports help take networks down. Fourth, inform your circle if there is a risk of impersonation: family, company, clients. Sometimes the fraud continues because the scammer exploits the stolen identity. Fifth, if there is significant financial loss or reputational damage, consult a lawyer to assess civil and criminal actions, bank claims for fraud-induced transfers, and platform liability if there were serious security failures or deceptive advertising. The routes vary by country, but there are more and more precedents for partial recovery when you act fast and document well.

The conclusion is clear: AI has not only created new scams, but it has also perfected the old ones until they are almost indistinguishable from reality. That is why defense no longer depends on “spotting what is fake,” but on changing habits: always verify through an alternative channel, distrust urgency, and remember that in 2025 any voice or video may be a mirage. AI digital literacy is no longer optional: it is basic self-protection.

The second category is deepfake video scams or fake ads. In 2024–2025, campaigns have exploded where supposed celebrities, politicians or business leaders recommend “miracle” investments, cryptocurrencies or financial products. Investors and consumers see videos with real faces saying things they never said. Recently, waves of deepfakes of public figures have been reported promoting investment frauds, precisely because people trust the authority of the person. Here the trap is psychological: if so-and-so says it, it must be serious. And it isn’t.

Third category: hyper-realistic, multilingual phishing. In the past, fraudulent emails were obvious because of spelling mistakes. Now, with generative AI, scammers write like a real company, in your language, with flawless style and personalized references. Europol and cybersecurity firms note that AI can generate perfect messages in seconds, adapted to each victim’s profile. A typical example: you receive an email supposedly from your bank or a major shopping platform saying there is suspicious activity, and you must “verify your identity”. The link takes you to a cloned website identical to the real one. AI has produced the site, the texts and the support chat. You enter your data. Account taken over.

Fourth category: cloned online stores and fake shopping ads. E-commerce fraud has multiplied because AI lets criminals create, in minutes, pages almost identical to well-known brands, with realistic photos, fake reviews and targeted advertising. In recent campaigns (especially during shopping seasons), waves of fake websites have been detected driven by deepfake ads or AI-generated messages. Example: you see on Instagram an offer from a familiar store with “50% off only today”. The ad looks real. The website is identical. You pay. Nothing arrives and your data is exposed.

Fifth category: sextortion and manipulation with fake images. AI allows criminals to create fake intimate photos or videos (or highly convincing edits) to blackmail victims. Although this threat is less “massive” than phishing, its emotional impact is enormous. The victim receives a message: “we have your images, pay or we will send them to your contacts.” Many times,the images are not real, but fear paralyzes. Regulators such as Ofcom in the UK have highlighted the growing presence of harmful deepfakes and related scams, with a high level of realism perceived by users.

The most dangerous thing about all these scams is that they have no “victim profile”. Older and younger people fall for them, executives and students alike. Because the decisive factor is not the victim’s intelligence but their emotional state and the urgency of the moment. AI is designed for that: to sound and look authentic precisely when you cannot verify.

What can you do to avoid falling for it? The golden rule is simple: distrust urgency. Every AI scam pushes you to act fast. If someone asks you for money “right now”, break the script. Breathe and verify through another channel. If it is supposedly your child, call their real number yourself. If it is your boss, write through your usual internal channel. If it is the bank, go into the official app, never through links. And if the message comes with video or voice, remember that today that no longer proves anything.

And if you have already fallen for it, what matters is to act calmly and quickly. First, stop the damage: block cards, change passwords, activate two-factor authentication, notify your bank or platform to freeze movements. Second, collect evidence: screenshots of emails, chats, numbers, links, audios, ads, transfer receipts. Third, report it: police or specialized prosecutors, and also the platforms where it happened WhatsApp, Instagram, TikTok, email provider, because their reports help take networks down. Fourth, inform your circle if there is a risk of impersonation: family, company, clients. Sometimes the fraud continues because the scammer exploits the stolen identity. Fifth, if there is significant financial loss or reputational damage, consult a lawyer to assess civil and criminal actions, bank claims for fraud-induced transfers, and platform liability if there were serious security failures or deceptive advertising. The routes vary by country, but there are more and more precedents for partial recovery when you act fast and document well.

The conclusion is clear: AI has not only created new scams, but it has also perfected the old ones until they are almost indistinguishable from reality. That is why defense no longer depends on “spotting what is fake,” but on changing habits: always verify through an alternative channel, distrust urgency, and remember that in 2025 any voice or video may be a mirage. AI digital literacy is no longer optional: it is basic self-protection.