Quote from Chukwuma on May 12, 2026, 11:18 am

AI-driven fraud is starting to look less like a consumer inconvenience and more like a financial-sector risk that investors may need to track closely. The US recorded the highest number of data compromises in 2025 since the Identity Theft Resource Center began tracking them in 2005, while Experian said 40% of the 5,000 data breaches it serviced last year were powered by AI. Experian also expects agentic AI, which uses multiple autonomous agents to pursue sophisticated goals with limited human oversight, to become the number one cause of data breaches this year. That could raise the stakes for banks, credit agencies, schools, health-care organizations and other data-heavy institutions, because these systems can scan the dark web for vulnerable Social Security numbers, impersonate different identities across multiple banks, and complete complex government loan forms far faster than older fraud operations.

The pressure is already visible across industries tied to sensitive data. Financial services recorded 739 data breaches in 2025, followed by health care at 534, professional services at 478, manufacturing at 299 and education at 188, according to the Identity Theft Resource Center. Identity theft reports to the Federal Trade Commission have climbed nearly 20% year over year, while TransUnion (NYSE:TRU) said more than $534 billion is lost to fraud globally each year. The bigger concern could be synthetic fraud, where scammers combine real and fake information to build realistic identities, open accounts, create small lines of credit, and possibly wait years before busting out by maxing out cards and withdrawing money. Tools such as FraudGPT can test large numbers of Social Security numbers in minutes, while AI can also help fraudsters create deepfake driver’s licenses and more convincing phishing emails.

The fight is now becoming an AI-versus-AI arms race. Anthropic is rolling out its Mythos model to a select group of companies to test products for vulnerabilities, while OpenAI is also shopping an equivalent model for companies to test. SEON said it uses more than 5,000 analysts to review transactions and assess fraud threats, while TransUnion uses AI for selfie liveness checks and can flag whether images are AI-generated. For investors, the takeaway is that fraud risk could possibly influence compliance spending, credit losses, customer trust and cybersecurity budgets across financial services and other sectors holding personal data. Consumers can freeze credit, use multi-factor authentication, rely on password vaults, check annual credit reports, avoid public Wi-Fi without a VPN, and file reports with the FTC and local police after fraud, but the larger investment story is that institutions may need stronger upstream defenses before stolen data turns into future financial losses.

https://finance.yahoo.com/sectors/technology/articles/ai-powered-breaches-hit-40-130926086.html