The AI Scams Playbook: Deepfake Doctors, Deepfake Wealth Advisors and More

How disinformation is monetized 

Scams become extremely effective when the fraudster is able to mimic or mirror the target, because humans trust identities that seem familiar. Deepfakes enable scammers and manipulators to exploit this confirmation bias at scale, creating synthetic identities replicating specific traits, demographic or social attributes to target multiple audiences at the same time. Same playbook, same framework, same script, different face. 

Motivations behind different scam and manipulation playbooks

Social media networks are flooded with deepfakes and AI manipulated content. By some estimates, more than 90% of content on the internet will be AI-generated by the end of 2026, and despite regulatory and industry efforts to clearly label synthetic content, malicious actors don’t play by the rules.

Sometimes AI scams are obvious – the end goal of the fraudster is to convince the target to part with money – either to purchase a product or service, make a donation, or to invest in a scheme. In some cases the product or service is actually legitimate, but the content used for engagement seems bizarre or unusual. In these situations it is likely that the scammer is leveraging the social network’s algorithm to maximize exposure as part of an affiliate sales or marketing program for a real offer. For legitimate brands that don’t vet their affiliates or promoters, these rogue actors can negatively affect their reputation. For more unscrupulous companies however, deepfakes also leave the door open to plausible deniability, putting distance between deepfaked promotional videos and the brand itself. 

In cases where the content is nonsensical or triggers extreme feelings of outrage, confusion, or disbelief, the channel owner is likely sharing ‘rage bait’ or ‘click bait’ with the intention of building the account up so it can be monetized through advertising, sold on as an account with a sizeable following, or eventually used to distribute political disinformation to a broad audience. We see this a lot with social media accounts that intermittently change focus or theme. 

Because emotionally triggering content is engaging, and social media algorithms promote content that is engaging, it doesn’t matter if the content elicits a positive or negative emotional response, just as long as it has an extreme emotional impact. In fact, content that triggers outrage is more likely to drive a viewer to comment. At this point the account owner then makes use of bot farms to keep engagement going in the comments by pushing more emotionally triggering material or argumentative responses through automated means. 

Examples of deepfake scams and weaponized narratives

Deepfake doctors and wellness influencers 

One of the most common attacks we see is in the health and wellness space, is on celebrity doctors like Karan Rangarajan (Dr Rajan) and Hilary Jones, where legitimate videos are manipulated by lipsyncs and deepfake audio tracks, or in more sophisticated attacks the medical professional is cloned as an AI-generated avatar, so their likeness can be used to promote so called ‘miracle cures’ for a variety of ailments. 

Other attacks include networks of fully synthetic doctors or wellness influencers promoting supplements, oils, or other cures, and a variety of apps, guides, and courses targeted at those struggling with mental health issues.      

Some of the more dangerous health advisor scams involve bizarre or unfounded claims – “you’ve been doing this thing wrong your whole life…” – which are intended to farm engagement and drive polarizing responses in the comments. In many cases, networks of deepfaked doctors and health advisors are deployed to mimic and appeal to a specific audience demographic to maximize engagement.   

Deepfake wealth advisors and investment influencers

Deepfaked celebrity endorsements for crypto scams are a very common attack. One of the earliest and most infamous examples of this type of scam is the Quantum AI crypto investment scheme, which used manipulated videos of Elon Musk to engage targets. 

The fraudsters used generative-AI to apply lipsyncs to videos of interviews and speeches given by Elon Musk. Combined with a deepfaked audio track, it appeared Musk was endorsing a crypto trading platform called Quantum AI. In this case the operational infrastructure was quite sophisticated and victims who signed up for the platform would be relentlessly targeted by the scammers over the phone and via email in a bid to get them to invest. 

The Quantum AI scam also heavily leveraged advertising networks to push clips of the manipulated videos to a broad audience quickly. 

Other public and political figures who have had their likeness used to promote bogus trading platforms include British politicians Nigel Farage and Rishi Sunak, and money saving expert Martin Lewis, as well as billionaire investors Stanley Druckenmiller and Warren Buffett. 

It’s not just investment scams that promise to double your money or hit you with fake ‘withdrawal’ fees however, many of the crypto scams aim to trick targets into activating malicious smart contracts that drain their crypto wallets, posing as trading platforms, decentralized exchanges, or trading bots. 

There are also AI scams that blur the lines around outright deception. In one example that claims to promote ‘wealth advice’, a YouTube channel hosts a synthetic protagonist with hundreds of thousands of subscribers and millions of views on videos that claim to expose a social code on how the ultra wealthy behave. The end-game is to sell you a $30 ebook of ‘investment advice’, which skirts the legalities of this being a scam. But while it’s not clear who is behind the channel, the scenarios the protagonist is in, involving high-end cars, yachts, planes, and luxury locations, are 100% synthetic. 

Deepfake sympathy scams and free software scams

We’ve written about these scams in more detail, here, but the main premise of this attack leverages fully synthetic avatars in association with an emotionally engaging narrative to convince targets to buy products that are either sub-standard or never arrive.   

A variant of this scam manipulates authentic videos of children with terminal illnesses (in some cases even after the patient has died) to beg for donations to pay for treatment. This has the additional impact of creating confusion and negative sentiment around legitimate donation drives, when videos appear featuring the same real person, but with a different name and a different story, perhaps manipulated to share a completely different message. 

Another version of this scheme involves a network of synthetic influencers who claim to give you access to premium software for free, by showing a ‘secret workaround’ or more commonly by asking you to install a piece of software. This software is malicious and creates a backdoor on the target’s computer allowing criminals to harvest sensitive data like passwords and financial info, as well as to drain crypto wallets. 

Deepfake food influencers and disinformation scams

The deepfake food influencer is a more unusual format for the scam in terms of subject matter, but it fits the attack framework perfectly. Synthetic protagonists drive engagement with bizarre, or wildly incorrect claims about certain foods. Just like with the deepfake health advisors, there are fears that some of these suggestions could be damaging to one’s health, and the aim here is either to boost engagement through emotional triggers and disinformation and get follower numbers up, or sell an ebook of ‘nutritional advice’ in the same vein as the wealth advice scam. 

In fact, the videos for both are so similar that it might be the same people behind the accounts. But it’s entirely possible the playbook for deepfake-driven scams has become so operationalized that any similarities are purely coincidental. 

Another variant of the food disinformation scam features networks of synthetic ‘ex-employees’ from big brand supermarkets claiming to reveal a behind-the-scenes controversy (typically food related) that got them fired. In most cases these videos send the target to a data harvesting site that promises a free gift card in exchange for your data. But the gift card offer is fake. 

Deepfake crisis and natural disaster footage

One of the more cynical uses of generative-AI is seen when natural disasters or other crises occur. Following the recent flash-floods in Nepal and Tibet, tens of deepfaked videos appeared alongside real footage of the event. The videos ranged from the realistic – showing bridges collapsing and towns being swept away, to the ridiculous – with a shark swimming in the floodwater. Some of the videos had millions of views and the main incentive here is monetization. The deepfaked video of the Nepal flash floods below has unrealistic physics and disappearing people, suggesting a low-effort attempt to farm engagement.

High profile events can quickly become viral moments on social media and cynical deepfake creators can capitalize on that momentum to drive engagement and followers. Social media platforms typically enable accounts for advertising when they reach 10,000 followers. So these events can become a source of revenue at the time, and boost follower numbers on the channel, making it attractive for sale or giving it organic reach which could then be used to promote other content. 

While not technically fraudulent or illegal, there are concerns that deepfake-driven disinformation can create confusion and panic around disasters, potentially making the situation worse. There are even concerns that deepfakes could hinder rescue operations. During the Nepal floods, a travel agency shared low-quality ID card photos of several missing Singaporean nationals, which were then used to generate high-fidelity deepfakes of the missing persons, sparking concerns that the synthetic imagery could create confusion. 

Deepfake conflict footage and strategic disinformation 

Our recent report, The Role of Deepfakes in Cognitive Warfare, covers this phenomenon in great detail and features an extensive analysis of the weaponization of narrative through a cognitive attack framework and the Narrative Kill Chain. 

While operationalized and matured throughout the Russia-Ukraine conflict, a new era of cognitive warfare has been ushered in by the democratization of synthetic media and the ability to systematically forge compelling deepfakes, along with a rapid decrease in the cost and effort required to do so. This allows actors, both state-sponsored and non-state, to achieve a high level of sophistication in cognitive attacks, and accelerates their adoption for future conflicts. 

Deepfake-driven disinformation has been witnessed to a lesser degree in the theaters of conflict around India-Pakistan, and Israel-Palestine, and since the start of 2026, deepfakes have been adopted as a cognitive weapon in the escalations between Iran, and Israel and the US.

One of the most illustrative patterns is the recurring wave of synthetic videos depicting high-impact attacks on iconic infrastructure in the United Arab Emirates, particularly the Burj Khalifa and Dubai International Airport. These videos are not designed to document reality, but to simulate highly plausible disruption scenarios within globally recognizable urban environments.

Their effectiveness is due to the choice of targets: these locations are not only physical assets but symbolic anchors of safety, modernity, and international connectivity. By associating synthetic attacks with recognizable landmarks, the content immediately activates the familiar identity stage of the Narrative Kill Chain, not through a human subject, but through a shared spatial identity that viewers instantly recognize and trust.

While the fears around deepfakes of crisis situations outlined above are focused on (unintentional) confusion and panic, the objective is very much intentional. Not to generate mass panic in a classical sense – that the attacks are really happening now, but that they might. The aim is to introduce persistent cognitive friction, a low-grade but continuous layer of doubt, unease, and risk anticipation. Over time, this can manifest in subtle but strategically relevant effects: reconsideration of travel decisions, hesitation in business engagement, shifts in relocation choices, or broader reputational erosion of the target environment. 

Deepfake authority scams, police and government impersonation

Police forces and governments around the world have recently issued warnings about the rapidly escalating threat represented by criminals using deepfakes to impersonate law enforcement and state authorities. 

Generative-AI can be leveraged by malicious actors to manipulate legitimate media or to create synthetic scenarios intended to mislead the public.  

The FBI recently flagged an escalation in a criminal scheme, where scammers are now deploying deepfake videos of senior FBI officials alongside spoofed Internet Crime Complaint Center (IC3) websites to defraud previous fraud victims a second time.

Deepfake-driven political disinformation

Political arenas around the world are becoming deepfake battlegrounds. Nation states, activists, and foreign agitators are all using gen-AI to sway public opinion and push agendas through disinformation. 

In India in July, Maharashtra Cyber Police had to register a first information report (FIR) after a deepfake video went viral, claiming to show Delhi Police Deputy Commissioner Sumit Jha announcing his resignation over the handling of recent student protests. Then as protests at Jantar Mantar continued to unfold, it became clear that synthetic media was helping shape a political crisis in real time. Delhi Police said thousands of fake posts, manipulated videos and misleading content were circulated on Facebook, Instagram, X and YouTube platforms during the protests.

In the US 2026 midterms, deepfakes have become routine campaign material, with multiple confirmed deepfake incidents across Senate races in Texas and Georgia. 

In Poland’s 2025 presidential election, AI-generated images appeared in four viral disinformation videos alleging voter fraud, none carrying AI-disclosure labels, while in both Buenos Aires’ and Ireland in 2025, deepfakes released before polls opened falsely claimed candidates had withdrawn from the race.

Meanwhile, in the UK, a genuinely novel case has appeared in the shape of a synthetic persona generated to deliver sustained political messaging, tempered by plausible deniability for the sponsoring party. An AI-generated character – a rapper called Danny Bones – with links to far-right Reform splinter party Advance UK repeatedly intervened in the Gorton and Denton by-election, marking the first time an AI character has been linked to a political party.

While it’s unclear if the stunt had any impact on voting, the character’s AI-generated music tracks achieved more than 375,000 Spotify and YouTube streams, and social media videos on YouTube Shorts, Instagram Reels and TikTok were viewed nearly three million times.

A new twist on deepfake romance scams 

Celebrity scams and romance scams are some of the oldest playbooks in the fraudsters toolkit. Scammers have long been manipulating media to hijack the likeness of a celebrity, while gen-AI has made it possible to create hybrid and fully synthetic avatars that help craft a compelling narrative or scenario in a bid to convince victims to send them money. 

In a new twist on the classic catfish playbook, scammers are creating high-quality deepfakes of real OnlyFans models and other sexual content creators in order to leverage their legitimate identity and web presence, in something similar to a romance scam. 

The fraudsters create fake social media profiles populated with stolen authentic content, then use deepfake avatars to communicate with targets and promise custom content or other exclusive interactions in exchange for transfer of funds. This attack is novel in that the scenario creates two victims – the target who is conned out of their money, and the original content creator, who loses potential revenue and has their reputation affected, or even finds themselves in danger from angry fans who have been ripped off by an imposter.  

Observed deepfake threat vectors in AI scams

AI Face Swap
Replaces a face in a photo or video with another face, creating the appearance that the target individual performed the actions or expressions of the source.

AI Lipsync
Alters mouth movements in a video to match different speech or languages, allowing fabricated statements, disinformation, and scams.

AI Generated Video
AI-generated human faces created by Generative Adversarial Networks, producing photorealistic identities that do not correspond to any real person. Or AI-created digital identities based on real people, capable of impersonation, deception, and fraud.

Synthetic Voice
AI-generated speech that mimics human tone, rhythm, and emotion. In the observed examples we see synthetic speech being deployed to relay the key messages in the videos.