Fakes look real. The intonation is familiar. The emotion is convincing. In such a situation, it is natural to react immediately, rather than stop and check the facts.
Read more LKL debutants attracted a defender who shone abroad
This is exactly where one of the most important problems of the new digital age emerges: images and sounds are no longer as reliable evidence as we once thought. Artificial intelligence (AI) tools already allow the creation of increasingly convincing voice, face, photo, and video fakes. This changes not only the technology market but also the very logic of trust.
For a long time, we followed a simple rule: rumors can be disbelieved, but if I saw or heard it myself, then it is real. In 2026, we can no longer rely solely on this rule.
The era of synthetic reality
Until recently, convincingly faking a human voice or face required much more technical skills, time, and resources. Today, such tools are becoming more accessible, cheaper, and easier to use.
This does not mean that every fake is perfect. However, even imperfectly created content can appear convincing when a person feels pressured to react immediately. Scammers often do not need a flawless result – a few seconds of a convincing voice, a sense of urgency, and emotional pressure are enough for them.
Stanford University’s “AI Index 2026” report allows this problem to be assessed in a broader context. The number of documented AI-related incidents per year has increased from 233 to 362. This does not mean that all these incidents were related to voice cloning or video fakes, but the data reveals a general trend – AI-related risks are becoming more frequent, diverse, and harder to manage.
At the same time, transparency among major AI model developers is decreasing. The Stanford report states that the average score of the Foundation Model Transparency Index dropped from 58 to 40 out of 100. The greatest lack of transparency remains in areas where the public and regulators especially need more clarity: training data, computing resources, and post-release monitoring of model impact.
This does not mean that the use of every specific voice cloning or video generation tool is impossible to trace. However, it reveals a broader problem: the most powerful AI tools are spreading faster than society and institutions can establish clear rules for their oversight, traceability, and accountability.
It is not the device that is hacked, but the human reaction
Classic fraud methods often relied on poorly prepared emails, suspicious links, or calls from unknown numbers. Such scams were not always easy to recognize, but they were often betrayed by clear warning signs.
AI changes this situation. Now scammers can target not only passwords or bank data but also a person’s emotional reaction. If you hear the voice of a child, mother, spouse, or boss during a call, you first react to its familiarity and perceived threat, and only then start to rationally assess the situation.
The US Federal Trade Commission has warned that scammers may only need a short audio clip found online, for example, from a video posted on social networks, to use AI to imitate a close person’s voice. The more people share audio and video content publicly, the more material can be used to create fakes.
It is important to understand that such scams work not because people are stupid or careless. They are effective because they exploit natural human reactions: care for loved ones, the desire to help, fear of being late, and trust in a familiar voice.
When a person hears “you need to act now,” it becomes harder to calmly and critically evaluate the situation. That is why a sense of urgency, a request to tell no one, and a demand to transfer money quickly are some of the most important warning signs.
The zero-trust principle – also in everyday life
The so-called zero-trust principle has long been applied in cybersecurity. Its essence is simple: even if information or its source seems familiar and reliable, it does not mean you can blindly trust them. Additional verification is needed.
Previously, this logic was mostly applied to company networks, employee accounts, and devices. Today, it is increasingly necessary to follow it in everyday life as well.
If you receive an unexpected call or message in which a close person asks you to transfer money, reveal a password or login code, or perform another unusual action, voice or video alone is no longer enough. You need to additionally verify the person’s identity in another way.
Read more From «TikTok» to books: a phenomenon born on social networks is changing reading habits
This can be very simple. End the conversation and call back the close person at the number you already have in your phone, or contact them by another usual method. You can ask a question to which only your close one knows the answer. Even better – agree in advance with your family on a safety phrase.
The safety phrase does not have to be complicated. It can be a simple word or sentence that you do not share on social networks and that only your closest people know. If someone calls in a panic and asks for money, the first question can be very simple: “What is our safety phrase?”
If there is no answer, the conversation should be ended and the situation checked in another way.
A public voice and face become increasingly sensitive information
Until now, privacy was mostly associated with passwords, personal codes, bank data, or home addresses. However, in the AI age, information that we previously could consider quite harmless is becoming increasingly important: voice, face, manner of speaking, or everyday life videos.
This does not mean we have to stop taking photos, filming, or sharing moments of our lives online. But it is worth understanding that publicly available audio and video recordings can be used not only for communication or preserving memories but also for creating fakes.
Particular caution should be exercised regarding publicly available audio and video content of children, seniors, leaders, and other public figures. The same applies to employees who have access to sensitive data or make financial decisions. The more such material is available online, the easier it is to create a convincing imitation.
This is not a call to fear everything. It is a call to recognize that digital hygiene is changing. Previously, we protected passwords. Now we increasingly have to think about our digital voice and face footprint.
How is trust changing?
The most important conclusion is not that nothing can be trusted anymore. Such an attitude would also be dangerous because society needs trust. Without it, neither relationships, business, nor public life are possible.
The real conclusion is different: today we must trust differently than before – not blindly, but after verification.
Previously, it was enough to see or hear. Now, more and more often, it is necessary to additionally verify what we see and hear. Was the information obtained from an official source? Can you call back? Does the request correspond to usual human behavior? Is there no pressure to act immediately? Is there no demand to transfer money, reveal passwords, or login codes? Can identity be confirmed in another way?
In companies, this means clear rules for financial transfers, instructions from managers, and sharing sensitive information. If an employee receives a voice message or video call in which a manager urgently asks to make a transfer, that alone should not be enough to confirm the authenticity of the instruction. Such a request should be additionally confirmed through an official channel and follow a clear procedure.
In families, this means pre-agreed safety phrases and a simple rule: if an unexpected call causes panic and there is a request to urgently transfer money, first check the situation and only then act.
A new stage of digital literacy
“Seeing is believing” was a useful rule when it was difficult to fake images and sounds. But AI changes our attitude towards what we see and hear. Today, a realistic-looking image or a familiar voice is not necessarily reliable evidence – it may just be another signal that needs to be verified.
This does not mean we have to live in constant fear. We just have to adapt to new digital security challenges – just as we once learned not to open suspicious links, not to share passwords, and not to enter bank data on untrusted websites.
AI has not destroyed trust – it has simply changed its rules. From now on, trusting only what we see or hear is no longer enough. It is increasingly important to be able to verify the source of information, to pause before reacting, and not to succumb to artificially created urgency.
