AI Scams: Gallup Report Reveals $68 Billion Fraud Epidemic

AI scams have officially transitioned from a theoretical future threat into an active, highly organized crisis. As artificial intelligence and machine learning technologies are integrated into commercial industries, transnational organized crime networks have quickly co-opted these same innovations to scale their illicit enterprises. The integration of generative AI tools—ranging from high-fidelity voice cloning software to hyper-realistic video deepfakes—has lowered the barrier to entry for digital fraud, allowing low-skilled criminals to launch incredibly convincing attacks. This digital shift has escalated consumer fraud from a localized nuisance to a multibillion-dollar national security threat. Criminal syndicates no longer rely on manual, easily detectable scripts; instead, they deploy automated, AI-driven workflows that can target thousands of unsuspecting victims simultaneously.

The rapid democratization of artificial intelligence tools over the past several years has provided legitimate businesses with unprecedented tools for efficiency. Unfortunately, it has done the same for bad actors. Historically, cybercriminals relied on relatively transparent phishing schemes characterized by broken English, obvious spelling errors, and generic greetings. With the advent of large language models (LLMs) and advanced synthetic media generation, these telltale signs of deception have virtually vanished.

Scammers now use advanced generative tools to draft perfectly articulated, contextually relevant emails, text messages, and direct messages on social media. By analyzing publicly available data and social media profiles, AI systems can automatically customize phishing messages for specific targets, a process known as hyper-personalized spear-phishing. Because these tools operate algorithmically, a single operator can launch thousands of unique, personalized scams in a fraction of the time it once took to draft a single manual lure. This represents a fundamental shift in the economics of cybercrime: the cost of executing highly targeted, persuasive social engineering attacks has effectively dropped to zero.

The Dawn of AI-Enabled Fraud

The integration of artificial intelligence into the standard criminal toolkit has also bypassed traditional geographical boundaries. Previously, a scammer operating from a foreign jurisdiction was often limited by language barriers and cultural nuances. Today, generative AI models can translate and localize persuasive text flawlessly in real-time. This allows foreign syndicates to converse with targets in colloquial English, matching regional dialects and cultural contexts with ease.

Furthermore, machine learning algorithms allow criminals to automate the processing of stolen credentials and personal data. Syndicates can ingest massive, leaked datasets from corporate breaches, feed them into analytical AI engines, and automatically identify the most lucrative targets. This systematic identification of high-value individuals—such as wealthy retirees, business executives, or individuals with known family vulnerabilities—has amplified the efficiency of modern phishing campaigns. The resulting landscape is one where attackers possess the scale of automated computer programs but retain the persuasive warmth of human-to-human communication.

Breaking Down the Numbers: Gallup and Stop Scams Alliance Survey

A groundbreaking joint survey conducted by Gallup and the Stop Scams Alliance, titled United States of Scams: The Financial and Emotional Fallout, has laid bare the staggering scale of this modern epidemic. The study, which surveyed 5,173 U.S. adults between January 8 and February 18, 2026, presents some of the most comprehensive quantitative data available on how deepfakes and automated social engineering are penetrating American households. The results highlight a terrifying trend: fraud is no longer just a series of isolated incidents, but a pervasive, systemic threat that touches every segment of the population.

The Human and Financial Cost

The survey revealed that approximately 6% of U.S. adults—equivalent to roughly 15 million citizens—were successfully scammed out of money over the last year alone. Beyond the raw number of victims, the cumulative financial devastation is historic. Americans lost an estimated $68 billion to fraud in 2025, translating to an average daily drain of $186 million from the U.S. consumer economy. This loss does not merely represent a minor line-item expense; it signifies the erosion of retirement portfolios, the liquidation of life savings, and severe psychological distress for millions of families. Additionally, the survey highlights a grim reality of lifetime vulnerability, finding that 24% of all U.S. adults have fallen victim to a scam at least once during their adult lives.

AI and Deepfakes: The Silent Accelerants

Perhaps the most alarming finding of the Gallup and Stop Scams Alliance study is that victims reported that 12% of successful scams directly involved AI or deepfakes. However, researchers and cybersecurity experts warn that this 12% figure is almost certainly an undercount. Because generative AI has reached a level of realism that makes it indistinguishable from human interaction, many victims remain completely unaware that they were interacting with synthetic media, cloned voices, or algorithmic chatbots rather than real individuals. As a result, the true penetration of AI-driven fraud inside the United States is likely far higher, acting as a silent accelerant to a rapidly worsening crisis.

Inside the “Industrialized Fraud Channel”

Fraud experts and security practitioners have warned that deepfake technology is no longer just a theoretical concept used for political disinformation; it has transitioned into an industrialized fraud channel. Criminal syndicates are leveraging automated APIs, open-source model repositories, and readily available software-as-a-service platforms to structure, execute, and optimize their operations at scale. By reducing the human labor required to draft emails, build fake investment portals, or make voice calls, AI enables small criminal cells to mimic the capacity of large-scale enterprises.

Voice Cloning and Impersonation Schemes

One of the most insidious methods utilized by modern scammers is AI-powered voice cloning. With only a three-second audio sample—often harvested from social media videos, public recordings, or voicemail greetings—malicious actors can generate a highly realistic synthetic copy of a target’s voice. This cloned voice is then used in urgent “grandparent scams” or kidnapping hoaxes, where scammers call elderly relatives claiming to be a child or grandchild in desperate need of immediate financial assistance due to an emergency or legal crisis.

The emotional urgency, combined with the undeniable familiarity of the cloned voice, bypasses the victim’s natural skepticism, leading to quick and irreversible financial transfers. This form of social engineering exploits core human biology: our brains are hardwired to react with urgency and protective instinct when we hear the voice of a loved one in distress. Scammers capitalize on this neurological shortcut, ensuring that the victim acts on impulse before logical verification can occur.

Deepfake Advertisements on Social Media

In addition to direct telephone impersonation, synthetic video advertising has exploded across major social media platforms. For example, tech giants like Meta Platforms Inc. have faced intense regulatory and public scrutiny after reports revealed scam advertisers were deploying AI-generated videos of prominent public figures to target vulnerable demographics.

These deepfake advertisements often feature synthesized versions of trusted political or media figures promoting fraudulent government grants, fake Medicare benefits, or high-yield, low-risk cryptocurrency investment schemes. The ease with which these platforms approve automated ad campaigns has allowed scammers to hijack the institutional credibility of established public figures to siphon money from seniors and lower-income communities.

Organized Cybercrime Operating at Fortune 500 Scale

The sheer magnitude of these cyber-enabled financial crimes has elevated them beyond the realm of petty theft. Today, they represent highly coordinated, transnational organized crime. “These guys aren’t called organized crime for nothing. They’re actually organized, and they’re using their organization to start attacking us with scale now to a tune of $68 billion, which is like the annual revenues of Delta Airlines. It’s like a Fortune 500 company. It’s huge,” Stop Scams Alliance founder and CEO Ken Westbrook told NBC News.

Westbrook, a 33-year veteran of the U.S. intelligence community who served at both the Central Intelligence Agency (CIA) and the Office of the Director of National Intelligence (ODNI), has long argued that these syndicates operate with the discipline, structured hierarchies, and performance metrics of legitimate global corporations. Many of these organizations operate out of specialized compounds in Southeast Asia, using forced labor, professional software developers, and psychological profiling teams to maximize the yield of every target. By approaching scamming as an industrialized pipeline, they run shift-based operations where employees are trained in social engineering tactics, provided with automated translation tools, and equipped with custom-built generative AI platforms to construct elaborate personas that build trust over months before executing the final theft.

Comparing the Scale: National Loss Statistics

To comprehend the societal impact of this threat, we must contextualize the Gallup and Stop Scams Alliance data alongside other leading economic indicators of consumer fraud. The $68 billion stolen from Americans is not a minor leakage of funds; it represents a major macro-economic drain that rivals major national industries and places a heavy burden on the public tax base. Understanding these metrics is vital for policymakers attempting to measure the threat and allocate resources to federal law enforcement.

Metric	Statistical Finding (2025)	Implications & Context
Total Financial Loss	$68 Billion	Comparable to the annual revenues of Delta Airlines; a Fortune 500 scale.
Daily Financial Loss	$186 Million	A continuous bleed of capital from consumers to transnational syndicates.
Annual Victim Count	~15 Million Adults (6%)	Millions of individuals suffering both financial and emotional distress.
AI or Deepfake Involvement	12% of Successful Scams	Likely undercounted due to victims’ lack of awareness about synthetic media.
Lifetime Victimization	24% of U.S. Adults	Nearly one in four Americans have been scammed during adulthood.

Left of Boom: Prevention and Policy Recommendations

To counter an adversary that operates with the agility and scale of a Fortune 500 enterprise, security strategies must evolve beyond simple consumer awareness campaigns. Ken Westbrook advocates for a “left of boom” philosophy—focusing on stopping scams before they ever reach a victim’s screen or phone. This requires systemic interventions at the structural level, blocking fraudulent traffic, fake advertisements, and synthetic media at the telecommunications and internet service provider tier.

In the intelligence community, “left of boom” refers to the period before an explosion or attack occurs; by focusing on this phase, the Stop Scams Alliance aims to disrupt the cyber infrastructure of international syndicates before they can make contact with vulnerable Americans.

Collaborative Public-Private Initiatives

Combatting AI-driven fraud requires an unprecedented level of cooperation between government regulators, financial institutions, telecommunications networks, and tech platforms. Financial institutions must implement real-time, behavioral anomaly detection systems that can flag unusual transfer requests, especially those initiated by elderly or vulnerable clients under the influence of social engineering. Furthermore, social media networks must adopt stricter verification standards for advertisers, ensuring that political and celebrity endorsements are not synthetically engineered deepfakes designed to target vulnerable communities.

Empowering Consumers with Advanced Provenance Tools

As synthetic audio and video become increasingly difficult to distinguish from genuine media, the development of cryptographic provenance standards is essential. Technologies such as those pioneered by the Coalition for Content Provenance and Authenticity (C2PA) offer a potential solution by embedding digital watermarks or cryptographic metadata into genuine media files. By verifying the origin and history of digital content, consumer-facing software can automatically alert users when an incoming media file, voice call, or video clip has been altered or entirely generated by artificial intelligence. This would provide a robust, algorithmic layer of protection that does not rely solely on the victim’s ability to spot a highly convincing synthetic fake.

Moving Forward in the Fight Against AI Fraud

The integration of artificial intelligence into organized crime has permanently altered the parameters of cybersecurity and financial protection. As the latest Gallup and Stop Scams Alliance survey demonstrates, the threat is no longer distant; it is active, highly organized, and growing in both financial scale and technical complexity. Mitigating a $68 billion global fraud ecosystem requires treating cyber-enabled scams not as an assortment of individual misfortunes, but as a coordinated threat to national security and economic stability. Only through systemic, structural interventions, rigorous cross-sector collaboration, and proactive technological defenses can society hope to tip the balance back in favor of consumers and dismantle the industrialized networks profiting from AI-enabled deception.