The NNA asked the Google Gemini AI app about the risks of AI fraud. Here's what it had to say:
“What is the current risk assessment to the financial services industry regarding artificial intelligence-fueled synthetic fraud?”
The following is an unedited response from Goodle Gemini generated on 4-3-2024:
The risk assessment for the financial services industry regarding AI-fueled synthetic fraud is very high. Here's a breakdown of the key points:
- Increased Difficulty in Detection: AI allows fraudsters to create highly realistic synthetic identities by combining real and fake data. This makes it harder for traditional fraud detection methods to identify them.
- Accessibility of AI Tools: Generative AI tools that create deepfakes and forge documents are becoming cheaper and easier to use, empowering even less sophisticated criminals.
- Potential Losses: Estimates suggest synthetic identity fraud could cause billions of dollars in losses by 2030 for financial institutions. [Editor’s Note: The Deloitte Center for Financial Services estimates potential losses at $23 billion]
Experts believe AI-generated fraud will likely worsen before preventative technology catches up. Here are some additional details you might find helpful:
- Synthetic fraud involves creating fake identities using a mix of real and fabricated information.
- Deepfakes are used to create realistic audio and video recordings of real people to bypass security checks.
Overall, the financial services industry faces a serious challenge from AI-powered synthetic fraud.