AI makes online fraud more credible—smart people should learn to recognize warning signs
Tekoäly teki verkkohuijauksista uskottavampia – viisas varautuu tunnistamaan hälytysmerkit
AI summary
AI makes online fraud more credible: in 2024, a finance employee at Hong Kong's Arup was defrauded of about $25 million in an AI-faked video conference. AI greatly lowers the barrier to forging audio and video, and can also quickly read source code to find vulnerabilities and generate attack code. Scholars at the University of Jyväskylä believe the next threat is "LLM grooming" that manipulates AI: feeding content into training data to induce models to repeat a specific narrative; the public should verify from multiple sources and not trust the surface appearance of audio and video.