AI systems used in fraud while defenses lag
Criminals are deploying AI systems including deepfake video, voice cloning, and automated messaging to conduct romance and investment scams at large scale across multiple countries. Financial losses have risen sharply, with the Federal Trade Commission reporting $2.1 billion in social media scam losses in 2025 and UK fraud reaching £629.3 million in the first half of that year, while fraud specialists warn existing defenses were not designed to counter AI-powered schemes.
What happened
Romance scams and investment fraud increasingly employ AI technologies including deepfake video, voice cloning, AI-generated messaging, and instant translation to conduct large-scale fraud across multiple jurisdictions and payment systems.
According to the Federal Trade Commission, social media users lost $2.1 billion to scams in 2025, with investment scams accounting for half of losses and romance scams rising 35% in the UK during the first half of 2025.
Reported impact
UK Finance reported criminals stole £629.3 million during the first half of 2025, with romance fraud victims losing more than £102 million across 10,784 reported cases during 2025.
Fraud specialists argue that AI-powered scams compress the cost, time, and manpower required to execute large-scale emotional manipulation campaigns, enabling criminal networks to scale operations beyond what human operators alone could sustain.
Index analysis
Romance scams now operate at industrial scale using AI to conduct simultaneous interactions across multiple platforms and jurisdictions, with operators increasingly resembling organized international criminal groups rather than isolated schemes.
Traditional anti-fraud systems were not designed to confront the personalization and industrialized social engineering enabled by AI, creating a gap between fraud innovation and regulatory or institutional response capabilities.
Governance controls
The discussion among fraud specialists and payments executives increasingly centers on fragmented intelligence systems, platform accountability, behavioral analytics, and operational resilience, suggesting existing fraud frameworks may be inadequate to address AI-driven scams.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.