Global searches for “AI identity fraud” has surged 3,826% in the past year, with searches reaching around 14,000 per month globally after a further 24% increase in the last three months.
This transformative rise highlights our current reality where fraudsters can use AI to produce increasingly realistic documents and manipulated identity material without the specialist skills previously required.
For CX leaders, this requires a strategic plan to effectively strengthen controls while ensuring verification remains accessible and proportionate to risk.
Ben O’Brien, Managing Director, ID Card Centre, told CX Today that the growing accessibility of AI is exposing the limitations of traditional ID checks.
“The real shift is that forging documents used to require skill; now it just requires access to a generator. The output quality has jumped while the human check hasn’t,” he said.
“A five-second glance was always a weak control, but it was proportionate when fakes were weak too.”
A Growing Threat to Trust
Today, generative AI has made identity fraud harder to solve by making convincing forgeries faster, cheaper, and easier to produce without specialist knowledge.
As a result, this gives more people the ability to create material that may be difficult to distinguish through a quick visual inspection, exposing a weakness in traditional verification processes.
“The real shift is that forging documents used to require skill; now it just requires access to a generator,” O’Brien explained.
“The output quality has jumped while the human check hasn’t.”
This problem is particularly acute in retail or entertainment environments where employees have only seconds to assess an ID while serving customers or processing applications.
Beyond enabling AI to appear age appropriate, synthetic identities can combine fabricated or stolen personal information to create identities that may appear legitimate across different checks.
“There is no real document to compare against, so ‘does this look fake’ instinct doesn’t even apply,” he emphasized.
This increasing risk has made customer-facing sectors particularly exposed, often involving large numbers of rapid, face-to-face checks, often carried out by junior employees working under social pressure to keep queues moving.
“Age-restricted retail and hospitality face short, high-volume, socially pressured checks by often junior staff,” he said.
“AI-generated IDs are specifically marketed for beating exactly this.”
In recruitment environments, remote hiring means employers may receive identity documents digitally and conduct interviews without physically meeting candidates, making it harder to establish if the document belongs to the candidate, whilst creating significant legal consequences with incorrect right-to-work documents.
Furthermore, healthcare, real estate, and education face similar challenges where remote interactions or time-pressured decisions are common.
Across sectors, current organization workflows are not meeting the increasingly sophisticated fraud environment, as they are still designed to operate for a low probability of threats.
Design Your Security Journey Around Risk
To tackle this ever-growing issue, organizations must introduce a risk-based model that adapts the level of intervention according to the likelihood of risk.
By building verification systems into the customer journey as a series of checks, this will allow businesses to keep routine interactions simple and apply additional checks when a higher level of risk is indicated.
This can reduce unnecessary friction for legitimate customers and give organizations a more consistent way to manage suspicious activity.
“People think security and speed are trade-offs,” O’Brien notes.
“They’re not, if you design it in layers; friction should scale with risk, not get spread evenly across every customer who walks through the door.”
CX leaders will therefore need to consider where verification occurs and how much information each stage needs before asking the customer to take another action.
For online interactions, document verification systems can analyze identification during capture and assess features that are difficult to reproduce accurately in a simple digital image.
O’Brien continued:
“Document authentication at capture, reading the actual security features, such as the chip, UV/IR and MRZ, rather than relying on a human eyeballing a card, is the single biggest upgrade a business can make.”
After this, additional controls such as liveness detection and biometric matching can be introduced where the risk warrants them, helping establish that a real person is present, and comparing that person against the identity associated with the verification process.
Furthermore, businesses can incorporate background information and behavioral signals to identify activity that falls outside expected patterns.
The role of customer-facing employees will therefore most likely change as these systems become more capable, introducing human review as an escalation mechanism for ambiguous or higher-risk cases.
By making verification more predictable, this gives employees clearer information about when intervention is actually necessary, ensuring stronger verification without turning every customer interaction into a security investigation.
