Financial Census 2026
Financial Census 2026
87% of finance professionals have ignored a claim they believed was fraudulent. And 67% say they might submit a dishonest claim themselves if everyone else was doing it.
Fraud does not always arrive as a sophisticated cyberattack or a convincing fake invoice. Sometimes it looks like a few extra miles on an expense claim. A receipt submitted twice. A company subscription used for personal reasons. It's small enough to explain away and common enough to ignore.
The Medius Financial Census 2026 reveals a growing tolerance for these everyday acts of financial dishonesty. We call them shallowfakes: small frauds that may appear harmless on their own but can create significant risk and loss when they become part of the culture.
The problem is not that finance teams can't recognize fraud. It's that too many people have started to accept it.
The fraud you fear isn't coming from outside—it's already in your building. Our 2026 Financial Census reveals how small acts of financial dishonesty are being tolerated, ignored, and even normalized across finance teams.
Finance professionals know dishonest behavior when they see it. But that does not always mean they act.
87% have ignored a small expense, reimbursement, or claim they believe was fraudulent.
Only 12% say they have never done so.
And the problem goes beyond looking the other way:
would be likely to submit a minor dishonest expense claim if it were common among coworkers.
say they would be very likely to do it.
would justify a small dishonest financial act if they felt underpaid or undervalued.
would round up an expense or mileage claim if they thought nobody would notice.
consider small expense exaggerations acceptable when the amount is minor, compared with just 33% who consider them unacceptable.
believe shallowfakes are already common in the workplace.
Together, the findings point to a cultural risk that traditional fraud policies may not be designed to address.
The line between a harmless shortcut and fraud may seem clear on paper. In practice, employees don't always agree.
Here's what finance executives see as some of the most common forms of workplace fraud:
Submitting an AI-edited or digitally altered receipt or invoice.
Using a vendor invoice template to create a ficticious claim.
Copying a legitimate invoice and changing payment details.
Slightly inflating hours on a timesheet.
Generating a realistic-looking invoice using AI tools.
Using a company subscription or account for personal use.
Rounding up mileage by a small amount on an expense claim.
Submitting a receipt twice.
What stands out is how low the percentages are. Even the most widely recognized behavior was identified as fraud by only around one-third of respondents.
That leaves plenty of room for questionable activity to be rationalized, overlooked or treated as somebody else’s problem.
Shallowfakes are not limited to padded mileage claims and duplicate receipts. Generative AI is making it easier to create altered documents, realistic invoices and convincing payment requests.
93% of finance professionals are concerned about AI-generated fraud over the next 12 months.
Alarmingly, generating a realistic-looking invoice with AI was recognized as workplace fraud by only 32% of respondents. That gap between concern and recognition could give increasingly sophisticated fraud attempts the chance to pass unnoticed.
Threat awareness also varies considerably by market. In the US, only 30% identified generating a realistic AI invoice as fraud, making it one of the least-recognized behaviors surveyed. In Sweden, it was the most recognized form of fraud at 40%.
Finance teams are using several approaches to identify suspicious invoices and manipulated documents:
use automated software to detect fraudulent invoices or manipulated documents.
use AI tools for fraud detection.
undergo manual review.
use third-party verification.
Finance leaders need intelligent controls that catch anomalies, as well as clear expectations for teams that make it difficult to dismiss any anomalies they see.
Most fraud programs focus on stopping bad actors. The Census suggests organizations also need to consider what happens when otherwise-trusted employees begin to excuse dishonest behavior.
A small exaggeration may not look serious enough to investigate. But when employees see coworkers getting away with it, the behavior can spread. What starts as an exception becomes common practice, and what becomes common practice becomes harder to challenge.
That is the shallowfake gap: the distance between knowing something is wrong and deciding it’s serious enough to stop.
Do not limit automated checks to high-value or unusual invoices. Small discrepancies can reveal patterns that manual reviews miss.
Use multistage approvals, separation of duties, and risk-based controls to make sure one person cannot create, approve, and release a questionable transaction.
Continuously check changes to bank accounts, addresses, and payment details. A legitimate-looking invoice should never be enough on its own.
Employees need a clear and confidential way to raise concerns without worrying that they will be ignored or penalized.
Show employees what modern invoice fraud, altered receipts, and AI-generated documents look like. Annual policy reminders are not enough.
Remove the grey areas. Define what is unacceptable, apply the rules consistently, and make sure minor amounts are not treated as permission to look away.
Use intelligent anomaly detection from Medius to identify potential fraud and duplicate payments before money leaves the business. It can flag unusual invoice amounts, supplier-detail changes, and patterns that deserve a closer look.
See how Medius helps protect
AP from invoice fraud.