Attacks using deepfakes to bypass digital identity checks grew 180% over the past year, according to a study by LexisNexis Risk Solutions. The firm projects that one in every 100 of the 100 billion identity verifications expected globally in 2026 will fail because of AI-generated forgeries. Fraudsters are concentrating on high-value, reusable documents—passports, driving licences and national ID cards from the United States, the United Kingdom, Germany and France—to seize accounts, authorise illicit payments and launder money.

In parallel, a series of controlled studies is mapping how generative AI reshapes human cognition. A British-American paper involving 1,222 participants, currently under peer review, found that using AI tools for arithmetic and reading-comprehension exercises improved immediate performance but reduced long-term retention and the capacity to persevere without assistance. A separate 2025 MIT study, which circulated widely, suggested that students who used generative AI to draft academic essays displayed weaker critical-thinking skills. Researchers at Carnegie Mellon and the French National Centre for Scientific Research (CNRS) describe a mechanism of “cognitive delegation”: the brain’s innate drive to conserve energy leads users to accept instant answers, weakening the neural connections that sustained effort would otherwise reinforce.

Viewed from technology developers, the response has been to embed so-called Socratic modes. OpenAI’s ChatGPT and Google’s Gemini now offer guided-learning settings that pose questions and hints rather than delivering full solutions. Microsoft confirmed it has built error-risk warnings and prompts into Copilot to keep users in an active, critical role. The company acknowledged that “the risk of excessive cognitive delegation is real, especially if AI is used to automate tasks that also serve to develop skills.”

In Mexico, regulators and private-sector security firms are pushing for wider adoption of biometric systems to counter a parallel wave of SIM-swapping and identity theft, with financial fraud complaints rising 32% year-on-year in early 2026. Analysts in Jakarta point to a deeper trust crisis: the “liar’s dividend” allows public figures to dismiss genuine evidence as deepfakes, while experts warn of an “infocalypse” in which societies lose faith in all audiovisual proof. The next factual milestone is the crossing of 100 billion identity checks this year, a volume that will test whether detection systems can keep pace with the sophistication of AI-generated deception.