Artificial Intelligence (AI) is evolving rapidly, but so are its risks. With the recent rollout of photo generation features in tools like ChatGPT, cybersecurity experts are raising serious concerns. The fear? AI-generated fake documents — including Aadhaar cards, passports, and KYC forms — that are almost impossible to detect with the naked eye.
New feature in ChatGPT sparks security concerns
OpenAI recently introduced a photo generation feature in ChatGPT, giving users the ability to create highly realistic images. While this has exciting use cases for creativity and productivity, it also opens a dangerous door. Experts now warn that cybercriminals can misuse this feature to generate convincing fake ID documents, leading to large-scale fraud.
KYC fraud threatens multiple sectors
AI-generated documents are not just a theoretical threat. The risk of KYC (Know Your Customer) fraud is especially high in sectors like banking, insurance, telecom, logistics, education, and healthcare. Using AI tools, scammers can easily forge ID proofs and commit serious financial crimes by impersonating real users or creating fake identities.
Current detection tools are not enough
Cybersecurity professionals have raised alarms that traditional fraud detection systems, like watermarking, facial recognition, and C2PA metadata, are no longer sufficient. Advanced anti-detection tools can now bypass these systems easily, making it harder for institutions to differentiate between real and fake documents.
As per experts, organizations need to invest in AI-specific deepfake detection systems that can identify manipulated or synthetic images from the moment of upload.
AI-driven cybercrime on the rise
The threat is not limited to fake documents. AI is increasingly being used to create deepfake videos and audio that look and sound extremely convincing. In 2024 alone, deepfake-related fraud has led to losses exceeding USD 6 billion globally. One Hong Kong-based multinational company suffered a massive USD 22.5 million loss due to such scams.
According to Ankush Tiwari, founder of cybersecurity startup pi-labs, these AI-powered threats are only going to grow. By 2028, experts predict that nearly 40 per cent of cyberattacks could involve deepfakes and social engineering tactics.