Generative AI: An Industrial Threat to Identity

⚡

Key Takeaways

1Generative AI and autonomous agents are accelerating identity theft, accounting for 40% of data breaches according to Experian.

2Fraudsters are using tools like FraudGPT to create deepfake identities and open fraudulent accounts.

3Global losses due to fraud exceed $534 billion, with an increase in the sophistication of attacks.

💡Why it matters — While AI facilitates fraud, it is also crucial for developing effective defenses against these growing threats.

Generative AI and Autonomous Agents: An Industrial Threat to Identity

Generative AI and autonomous agents are transforming identity theft into a large-scale operation in the United States. According to the credit bureau Experian, 40% of the data breaches examined last year involved the use of AI.

Fraudsters are exploiting tools like FraudGPT to identify valid social security numbers and create deepfake identities, thereby opening fraudulent accounts. Global annual losses due to these frauds now reach $534 billion.

To counter this threat, companies are implementing automated presence checks and AI-based risk scoring systems. Individuals are encouraged to use credit freezes, multi-factor authentication, access keys, and VPNs.

A Revealing Bloomberg Investigation

An investigation conducted by Bloomberg sheds light on the impact of generative AI and autonomous agents on identity theft in the United States. From searching for social security numbers on the darknet to deepfake driver's licenses, the methods are varied.

The case of Jennah Haque, a reporter at Bloomberg, illustrates this issue well. She received a package from the Ultimate Medical Academy in Tampa, which included a men's polo shirt in size XL, despite never having applied. Someone had submitted 13 college applications and several financial aid requests in her name, potentially unlocking over $50,000 in student loans. The name, date of birth, address, and social security number matched perfectly. The only clue: a high school in Alabama where Jennah Haque had never set foot.

A financial aid employee in the U.S. told Haque that the volume of college applications submitted in such a short time would be nearly impossible without the assistance of AI.

The Scale of the Phenomenon

AI-driven identity theft has reached unprecedented levels, as shown by figures from the Identity Theft Resource Center (ITRC) for 2025, indicating the highest number of data compromises since tracking began in 2005.

Michael Bruemmer, vice president of consumer protection at Experian, reports that 40% of the 5,000 data breaches handled last year involved AI. He predicts that by 2026, agentic AI will become the primary driver of these attacks.

Sophisticated Tools and Varied Methods

AI tools are already highly advanced. FraudGPT, for example, can test hundreds of thousands of social security numbers in minutes to find a valid combination. Sub-agents scour the darknet for exploitable personal data, contact multiple banks under different identities, and automatically fill out complex government forms.

Naureen Ali, fraud manager at TransUnion, describes a common fraud scheme: fraudsters open small lines of credit at local banks, then larger lines with institutional lenders, using deepfake driver's licenses for identity verification.

Experts' Response

Experts agree that the best defense against this wave of fraud is the use of AI itself. TransUnion employs automated presence checks to detect AI-generated selfies, while SEON analyzes transactions using proprietary risk scores.

For individuals, standard advice remains valid: credit freeze, multi-factor authentication, access keys, and avoiding public Wi-Fi without a VPN.