LinkedIn user prompts ChatGPT to create fake passport

A LinkedIn user has exposed how a ChatGPT-generated fake passport bypasses digital identity checks, revealing a major vulnerability in the verification process.

Imagine if you could generate a fake ID passport in five minutes to plan your Breaking Bad-style getaway.

Well, one user claims to have managed to do this using the new Chat GPT-4o without any hitches.

Venture capitalist and tech entrepreneur Boris Musielak, raised a pertinent question on LinkedIn when he generated a fake passport at breakneck speed.

The debate rages on about whether big corporations know your customer (KYC) verification process is still relevant or if it’s becoming obsolete.

Passport to fraud

As it stands, mainstream AI generally doesn’t have a great reputation when it comes to generating fake documents.

Common weaknesses include poor machine-readable formatting and layout consistencies.

Musielak’s fake ID upload, however, is a seamless and credible passport clone, and as he posted the copy, many LinkedIn users discussed the issue.

Many users pointed out that this is a much faster way than using Adobe Photoshop. One user discussed the implications of AI cloning images, videos, and voices to commit major fraud.

Musielak conceded to some opponents that the image might not be perfect and perhaps wouldn’t pass a chip scan, as is the case with some Smart ID scans, such as those used for immigration purposes online.

When approached for comment Musielak told Cybernews: “The risks of fraud, stealing of your personal data for credit check purposes just significantly increased as it's now easier than ever for fraudsters to try such malice practices at scale.”

Some online banks like Revolut or cryptocurrency exchanges like Binance simply ask their users to upload a selfie.

Deepfake images could well bypass such systems, with fraud and illicit transactions becoming rampant as a result.

Therefore, near-field communication (NFC) is a solution that should prevail in this case. This form of validating ID would be applied to devices validating a document.

Institutions like Fraud.net and Sensity are currently developing AI to try to repel the threat posed by Deepfakes.

"If your "KYC" depends solely on pictures or videos, you should upgrade immediately to an eID and NFC / driven solution. Digital identities are immune to genAI fraud as opposed to the pictures or videos of the analog identities” advises Musielak.

Since Musielak’s uploading of his fake passport, I myself tried to do the same, but around 16 hours after the original LinkedIn post, ChatGPT refused to cooperate, saying:

“I can't create or generate any kind of fake or unofficial ID cards using personal or altered details.”