Life is at times a dichotomy, at times paradoxical and most of the time ironical. If life is so, most of the time, most things are some variant of it. And Artificial Intelligence is no different. While most of the world finds it as a magical ready-made tool to do most of their personal and a substantial part of the official work with the least amount of drudgery and very little headache. It’s just another way of delivering smarter, more cohesive, answers and results the way you want, a LLM personified in a conversational mode. App at the end of it. That there can be AI fraud is nowhere on our mental horizon.
India is still having its tryst with Digital Arrest, not being able to make a dent with its half-baked efforts. It would be unimaginable for our governments and even the tech titans to imagine that the AI fraud crisis has arrived. They are not the only ones. The global poster boy of the AI revolution also feels the same. Sam Altman had recently warned that AI-powered fraud is coming “very soon”, and it will break the systems we rely on to verify identity. Haywood Talcove, the CEO of the Government Group at LexisNexis Risk Solutions, has a totally different take.
In the words of Haywood Talcove “I’m a cybersecurity CEO who advises over 9000 agencies and Sam Altman is wrong that the the IA fraud crisis is coming, it’s already here.” This is the wide chasm between the creators / developers and the ones who have to face the negative onslaught of these double edged technologies. Talcove works with federal and state agencies on fraud prevention. While talking of the US, he says, AI generated fraud is siphoning millions from public benefit systems, disaster relief funds, and unemployment programs.
There are organised criminal networks proficient in deepfakes, synthetic identities, and large language models very clearly outpace and outbeat the outdated fraud defenses, which includes easily spoofed single-layer tools like facial recognition, and as expected they are winning. Pandemic saw billions swindled in unemployment benefits. AI-generated fake identities, voice clones, and forged documents overwhelmed systems that weren’t built to detect them. Today the tactics are more advanced and fully automated. The Small Business Administration now estimates that nearly 200 billion was stolen from the pandemic-era unemployment programs. As AI evolves, the scale and sophistication of these attacks will increase exponentially.
AI WITHOUT GUARDRAILS FALLOUT WOULD BE VERY DIFFICULT TO HANDLE OR CONTAIN.
Sanjay Sahay
Have a nice evening.