AI‑driven cyber fraud and legal hallucinations raise global security concerns

Artificial intelligence is being weaponised in two distinct but related ways. In the cyber realm, AI tools are automating fraud schemes that have already cost India over ₹52,976 crore ($5.8 billion) in six years, with a single British engineering firm losing $25 million to synthetic‑voice phishing. The World Economic Forum notes that 73 % of respondents reported personal exposure to cyber‑enabled fraud in 2025, and AI‑generated deepfakes are deepening deception across markets and politics.

In the legal sector, the UK Crown Prosecution Service unintentionally cited two fictitious legal authorities in High Court filings after they were produced by generative AI. The error was caught before the hearing and caused no adverse outcome, but it prompted an internal review of nearly 80 related cases and highlighted the risk of AI hallucinations embedding false information into official records. Both incidents underscore the need for stronger human oversight, verification processes, and coordinated international safeguards to prevent AI from undermining security and the rule of law.