Governments take action against the proliferation of deepfakes
Malaysia and Indonesia made headlines this week by restricting access to Grok, the AI chatbot developed by Elon Musk’s xAI. The move comes after authorities raised serious concerns about the platform being used to generate sexually explicit and non-consensual images. California Attorney General Rob Bonta announced a similar investigation, confirming that his office was looking into several reports involving sexualized images of real people.
Bonta’s statement was quite direct. He said the material, which shows women and children naked and in explicit situations, has been used to harass people on the internet. He urged xAI to take immediate action. But I think the real problem here goes beyond a single company or platform.
The evolution of deepfake technology
What’s different about new deepfake tools is their dynamic responsiveness. Unlike earlier versions which were more static, these tools can respond to prompts in real time. They reproduce natural facial movements with convincing accuracy: head movements, blinking, smiling, and all seeming authentic. Synchronization between speech and facial expressions has improved significantly.
This advancement creates a problem for verification systems that rely on these basic controls. Asking someone to blink or turn their head during a video check might no longer work. Technology has simply become too good at imitating these natural movements.
Implications for centralized exchanges
For crypto platforms, this presents a real challenge. Most centralized exchanges use some form of visual identity verification during onboarding. This is part of their Know Your Customer (KYC) requirements. The process usually involves users submitting photos or videos of themselves with identifying documents.
But if deepfakes manage to convincingly replicate these verification steps, the entire system becomes vulnerable. The financial impact is no longer theoretical. Industry observers have noticed AI-generated images and videos appearing in insurance claims and legal disputes. Crypto platforms, with their global reach and often automated onboarding processes, could become attractive targets.
The need for adaptive security measures
The real question is not whether users of centralized exchanges should worry – they probably should, at least a little. The most pressing question is how platforms will adapt. Trust based solely on visual verification may not be enough in the future.
Crypto exchanges face the challenge of updating their security measures before technology overtakes their protections. It’s not just about adding extra layers of verification, although that could help. This is about fundamentally rethinking how identity verification works in a world where visual evidence can be fabricated.
Some platforms may need to incorporate additional verification methods. Perhaps behavioral analysis, device fingerprinting, or more sophisticated biometric controls. The problem is that each additional layer adds friction to the user experience, which exchanges generally try to minimize.
There is also the regulatory aspect to consider. As governments around the world pass laws against deepfakes, exchanges will need to ensure their compliance measures keep pace with technological and legal developments.
What seems clear is that the debate around identity verification needs to evolve. Old methods may not hold up to new technologies. And for centralized exchanges, that means re-evaluating their entire approach to user verification and security.
