When employees can’t spot a deepfake it becomes every company’s problem

Americans struggle to distinguish real from AI-generated content

Most people in the United States can’t tell what’s real and what’s fake online anymore. That’s a foundational problem for any company relying on digital identity verification. The challenge here isn’t awareness. It’s capability. When users can’t tell if an image or video is authentic, the entire process of verifying identity in sectors like banking, ecommerce, or enterprise access control becomes fragile. Deepfakes now make it easy for bad actors to impersonate real individuals, bypass system checks, and commit fraud at scale.

This is about protecting digital infrastructure. Systems built for a pre-AI era need to evolve fast. Identity verification, once a compliance checkbox, must now become a core capability. When technology shifts this dramatically, the companies that move early and rebuild their frameworks with automated verification tools will define the new standard of trust online.

A 2026 Veriff–Kantar study tested deepfake detection on 3,000 people across the U.S., U.K., and Brazil. Americans scored just 0.07 on a scale where 0 equals random guessing. That statistic confirms that human eyesight and intuition are no longer reliable defenses. Machines are beating humans at generating realism, so the only feasible response is machine-driven verification.

Ira Bondar-Mucci, fraud platform lead at Veriff, summed it up clearly: “Now that AI-generated content is becoming indistinguishable from reality, the human eye alone is no longer a reliable line of defense.” Businesses and policymakers need to treat identity verification not as optional tech, but as essential infrastructure. Those who act now will create a real competitive moat in digital trust.

The U.S. exhibits a pronounced deepfake awareness gap despite its leadership in AI development

The United States leads the world in developing generative AI but ranks behind in recognizing its risks. Awareness of deepfakes among U.S. adults sits at 63%, compared with 74% in the U.K. and 67% in Brazil. That gap directly increases the probability of fraud and misinformation spreading unchecked. When consumers don’t understand what a deepfake is, they don’t question authenticity. They trust too easily. For businesses, this misplaced trust becomes a systemic weakness.

Many American consumers have historically focused on privacy and data protection, not authenticity. That mindset made sense before AI fakes became indistinguishable from real content. Today, that focus must expand. Lack of awareness doesn’t lower risk, it amplifies it. It gives attackers an edge and forces companies to carry the cost when verification fails.

Executives should see this not as a communications problem but as a structural one. Awareness campaigns alone won’t fix it; awareness needs to be built into the design of digital systems. Education should run parallel to automation. Artificial intelligence can detect synthetic content faster and more accurately than any team of human reviewers, but end users still need to understand why verification cues matter.

As Ira Bondar-Mucci noted, “There’s a paradox at play… if you don’t know what a deepfake is, you’re far less likely to pause and verify whether you’ve encountered one.” The next generation of digital systems, and policies, must close this gap. Awareness and automation must evolve together, or American users will continue to be the weakest link in their own digital security chain.

Human detection of deepfakes is unreliable across different content types

Human intuition has reached its limit when it comes to detecting synthetic media. The research shows that people perform almost randomly when asked to judge whether visual content is real or fake. Even when presented with side-by-side comparisons, most participants split evenly between believing or doubting what they saw. Video, in particular, proved most difficult, participants frequently judged fake videos as real and genuine ones as false. This means traditional visual verification is no longer a dependable security measure.

For businesses, the implication is immediate. Any process dependent on human judgment for visual validation, such as onboarding, access control, or content moderation, is inherently exposed to error. Automated systems must take on that responsibility. AI-based verification can analyze patterns and inconsistencies at the pixel or metadata level that humans cannot perceive. The role of people in the verification chain should shift from judgment to oversight, focused on reviewing the rare and complex cases automation cannot resolve with confidence.

Executives should see this as both an operational risk and an opportunity. The cost of relying on human review is no longer in staffing, it’s in exposure. As synthetic media technology improves, every manual verification process becomes a potential entry point for fraud or misinformation. Companies that act early to replace manual decisions with automation will not only reduce risk but also strengthen digital trust across their platforms and transactions.

The Veriff–Kantar findings make clear that even trained eyes cannot consistently identify deepfakes. Human accuracy hovers near random levels, confirming that the standard of “see and decide” has expired. Verification must evolve to meet a reality where what appears authentic may not be.

Overconfidence in deepfake detection amplifies consumer vulnerability

Confidence without capability has become one of the biggest liabilities in digital security. Around half of U.S. respondents in the survey said they were confident in spotting deepfakes, but most of them performed poorly. A smaller subset, roughly 7%—stood out for a troubling combination of overconfidence and inaccuracy. They rarely verify what they see, even when confronted with potentially manipulated content. In practical terms, this group represents millions of accounts vulnerable to exploitation.

This confidence problem extends beyond individual users. Businesses that depend on customers to self-verify or employees to manually assess visual content are adopting that same vulnerability by extension. When an organization trusts human perception over technological detection, it internalizes human bias, error, and misplaced certainty. Fraudsters understand this and design attacks to exploit those very points of human weakness.

Leaders should approach this issue strategically. Overconfidence is not a user-interface problem, it’s a behavioral one. Automation mitigates it by removing subjectivity from the process. Systems built to detect synthetic content before a human decision is needed can drastically lower risk. These systems allow teams to focus on response and mitigation, not detection. The less an operation depends on user judgment, the fewer chances there are for confidence to override caution.

As Ira Bondar-Mucci from Veriff stated, “When people believe they can’t be fooled, they stop looking for the signs. That’s precisely when they’re most vulnerable.” That’s the reality executives must account for when designing secure digital systems. The goal is not to eliminate human involvement but to ensure that humans are never the first or only line of defense.

Despite high concern over deepfakes, americans tend to delegate detection responsibilities to digital platforms

Americans are deeply concerned about the rise of deepfakes but tend to assume that platforms will handle the threat. The Veriff–Kantar data shows that 79% of U.S. respondents are “rather” or “extremely” concerned about being targeted by impersonation or fraud. Yet many of those same people trust social media and digital platforms to identify AI-generated content for them. The result is overconfidence in platform safeguards and a decline in personal vigilance.

For businesses, this misalignment between concern and action presents a risk that extends well beyond individual users. A public that trusts too easily creates fertile ground for fraud and misinformation to spread. Companies that manage high-value transactions or digital interactions cannot rely on external platforms to secure their users’ identities. Security and authenticity verification must be embedded into the company’s own systems, not outsourced to public trust.

Executives should view this as a moment to lead rather than delegate responsibility. Strong internal verification protocols send a clear message to customers that authenticity is being treated as a core business asset, not a side function. Companies that set this standard early will gain a competitive edge, both in public reputation and regulatory alignment. As global regulations tighten around synthetic media and fraud prevention, those already invested in detection and verification infrastructure will be better positioned to scale without disruption.

Ira Bondar-Mucci from Veriff underscored this dynamic by warning that the gap between consumers’ perceived safety and actual protection “is exactly where fraud thrives.” That statement reflects a key truth: awareness alone is not protection. Security cannot depend on confidence; it must depend on verifiable control.

The business case for automated, AI-powered identity verification is imperative

Relying on human judgment to confirm identity in an AI-saturated world no longer works. The problem isn’t that people lack awareness, it’s that no human can consistently detect synthetic content dense with algorithmic precision. This makes automated verification technologies not just a smart investment but a strategic necessity. Systems need to identify synthetic media before a human decision occurs, reducing both error and exposure across the entire business process.

The companies that integrate automated, AI-driven verification at every point of interaction will gain a lasting advantage. These systems can operate continuously and objectively, detecting deepfakes, forged identities, or synthetic data long before fraud can take root. Human oversight still has a role but only after AI systems have filtered out known risks. This division of labor ensures resources are spent on resolution rather than detection.

From an executive standpoint, automation is more than a technology choice, it’s an infrastructure decision. Deepfakes expose the flaws in legacy authentication systems, revealing that digital trust now requires proactive design, not reactive compliance. Businesses that continue to rely on manual verification or user self-attestation are embedding doubt directly into their systems. Those that automate will remove that weakness entirely.

Ira Bondar-Mucci summarized the shift in simple terms: “Seeing is no longer believing.” That statement captures the new operational reality. Companies must now build systems that validate truth independently of human sight. Those who adapt quickly will earn, and maintain, the trust needed to thrive as synthetic media reshapes the digital economy.

Key highlights

• Human perception is no longer a defense against deepfakes: Most Americans can’t distinguish AI-generated content from real visuals, leaving businesses that depend on visual identity checks exposed. Leaders should shift verification from manual review to automated, AI-driven systems to secure trust and authenticity.
• Awareness gaps magnify digital risk: Only 63% of U.S. adults understand what deepfakes are, despite the country’s leadership in AI. Executives should back both awareness initiatives and embedded verification systems to close this vulnerability and strengthen user vigilance.
• Human-based verification is unreliable and inefficient: People incorrectly assess real and fake content across all formats, even in direct comparisons. Companies should transition to automated detection technologies that identify manipulation at the data level before human review is required.
• Overconfidence increases exposure to fraud: Around half of U.S. users wrongly believe they can spot deepfakes, with 7% being particularly inaccurate yet certain. Leaders should eliminate dependence on user judgment by integrating automated validation tools that overcome human bias and overconfidence.
• Dependence on platforms undermines security: Americans trust tech and social platforms to manage deepfake threats, reducing their own awareness. Businesses should internalize identity protection measures to maintain control over digital security rather than relying on third parties.
• Automated, AI-powered verification is now fundamental infrastructure: Manual or user-driven identity checks embed risk directly into operations. Executives should prioritize AI-led verification frameworks that authenticate truth independently of human perception, ensuring scalable and durable digital trust.