The recent AI Festival in Ciudad Juárez, Mexico, served as a stark reminder that the rapid evolution of artificial intelligence is fundamentally altering our perception of reality and, consequently, the landscape of digital security. What was once the stuff of science fiction-a video call participant who isn’t who they claim to be, or a voice message from a loved one that is, in fact, an AI-generated fabrication-is now a tangible threat. Magaly Masís, an expert in computer law, cybersecurity, and digital forensics, brought these chilling scenarios to the forefront, challenging attendees to reconsider their trust in what they see and hear online.
The Unsettling Reality of Synthetic Content
Masís’s presentation, aptly titled “If you can no longer believe what you see, what will you believe? How to distinguish reality and protect yourself in the age of AI,” underscored a critical shift in digital interaction. The core premise of trusting visual and auditory cues is being eroded by the sophistication of AI-generated content. Through real-world examples and live demonstrations, Masís vividly illustrated how synthetic media can be weaponized for identity theft and elaborate deception.
A pivotal moment in the session involved the playback of an artificially generated video featuring Masís herself, making statements she had never uttered. The expert then polled the audience, asking who had identified the video as a simulation. The exercise served as a powerful testament to the limitations of human perception; the human eye is often no longer sufficient to discern the authenticity of an image or video, especially as AI generation tools continue to refine their craft.
The Convergence of AI and Social Engineering: A Potent Threat
The danger escalates exponentially when these advanced AI technologies are combined with social engineering tactics. Masís recounted a chilling 2024 incident where an employee was ensnared in a video call with individuals impersonating company executives. This interaction ultimately led to fraudulent transfers totaling approximately US$25.6 million. This case, Masís explained, exemplifies how a fabricated digital identity can become the genesis of a sophisticated fraud operation.
Voice cloning presents another formidable frontier of risk. Masís shared a personal anecdote of receiving a late-night call from someone attempting to convince her of an urgent situation involving her son. While the conversation was meticulously designed to induce panic, subtle inconsistencies allowed her to detect the deception. This personal experience highlighted how AI can amplify social engineering techniques that exploit human emotions rather than technological vulnerabilities.
Navigating the AI Deception Landscape: A Call for Vigilance
In response to this evolving threat, Masís proposed a simple yet effective rule for confronting unexpected requests: pause, analyze, reconfirm, and act. The pause breaks the sense of urgency often employed by attackers; analysis involves scrutinizing the sender and their motives; reconfirmation necessitates using a secondary channel or mechanism to verify identity; and action should only proceed after this thorough verification process is complete.
This recommendation extends to organizations as well. Masís stressed that AI adoption must be coupled with robust governance mechanisms, comprehensive inventories of AI systems in use, proactive risk management, stringent controls for sensitive transactions, and regular simulation exercises. The objective, she argued, is not merely to train employees on phishing, but to prepare companies for scenarios where an ostensibly legitimate call, video, or instruction could have been generated or manipulated by AI.
The specialist also advocated for integrating these practices into family life. One key recommendation was to establish
emergency keywords
that family members can use to quickly verify identities during suspicious calls or video chats. In the corporate sphere, she emphasized the importance of maintaining dual approval processes for payments and utilizing verification mechanisms that operate independently of the original communication channel.
The Ever-Evolving Challenge: Staying Ahead of AI
However, the challenge is dynamic and constantly evolving. The tell-tale signs currently used to identify AI-generated content-such as unusual movements, blurry contours, shadows, visual inconsistencies, or unnatural pauses in speech-may soon become obsolete as AI models continue to improve. Therefore, Masís insisted that the solution cannot solely rely on recognizing the imperfections of a technology that is in perpetual advancement.
The central question posed during the conference ultimately shifted from perception to verification: it’s no longer just about questioning whether we can believe what we see, but about understanding what mechanisms we possess to verify it. In an environment where artificial intelligence can produce increasingly convincing imitations of people, voices, and images, human discernment, rigorous verification, and robust governance are becoming paramount in the security strategies of both businesses and individual users.
The AI Festival in Ciudad Juárez is part of Connecta B2B’s 2026 tour, an initiative aimed at bringing the conversation about applied artificial intelligence in business to various markets across Latin America and the Caribbean. The Ciudad Juárez edition featured a diverse agenda, covering topics ranging from productivity and AI agents to enterprise implementation, manufacturing, sales, and digital security.
