As enterprises race to bolt AI onto every business process, security leaders are being forced to answer a harder question than "should we adopt it" — it's "who's accountable when it goes wrong." To unpack this, The Cyber Express sat down with Harsha Reddy, Head of Information Security at Veterinary Emergency Group (VEG).
With nearly two decades in security leadership — including senior roles at Lixil and American Standard before joining VEG — Harsha brings a practitioner's view of where AI is genuinely changing the CISO's job, and where it's mostly just hype and shadow adoption.
Watch the Full Interview:
Harsha Reddy Explains Why AI Will Replace Tasks, Not Defenders
Harsha pushes back on the narrative that AI will hollow out security teams, pointing to Gartner research showing that while most fields are projected to lose jobs to AI, cybersecurity is expected to gain them. In his view, the technology is mainly absorbing the "toil" — log review, alert triage, evidence gathering — that keeps analysts from actually defending.
“It's the analyst who refuses to use AI that will get replaced by an analyst who uses it,” he says.
On adoption, Reddy points to a 2024 Microsoft-LinkedIn survey in which most executives called AI critical to their business, yet a majority had no formal plan and most had employees already bringing in their own tools. That gap, he argues, is why so many organizations are now dealing with AI-related data leaks. "Many organizations started onboarding AI like software when they should be onboarding it like staff." His fix isn't more restrictions — blocking AI just pushes it into the shadows — but guardrails, an internal AI enablement committee, and measuring actual business value instead of token consumption.
The conversation also digs into deepfake-driven fraud, why training employees to spot deepfakes is “a losing bet” at machine speed, and how he decides when to greenlight a new AI tool versus telling a business unit “not yet.”
The conversation closes with our newly introduced rapid-fire round "Express Shots" — Claude vs. ChatGPT, passkeys vs. passwords, and Reddy's prediction for the biggest cybersecurity threat of 2030.
Read the latest updates Firewall Daily news and insights on The Cyber Expres, your trusted source for cybersecurity and information technology updates."
Seu telefone toca, você atende e diz “alô”. Do outro lado: silêncio total. Ninguém responde e a chamada é encerrada abruptamente. Se você ainda não usa [placeholder WhoCalls]bloqueadores de chamadas de spam[/placeholder], é quase certo que já tenha se deparado com essa situação.
Na maioria dos casos, trata-se de chamadas de golpistas. Hoje, explicamos por que essas chamadas acontecem, o que os autores da chamada querem de você e como se proteger. O mais importante é que vamos analisar se você realmente precisa se preocupar em se proteger contra esse tipo de ligação.
Quem está ligando?
Não são apenas golpistas do outro lado da linha: robôs, operadores legítimos de centrais de atendimento e pessoas comuns também fazem essas ligações. Vamos analisar cada tipo de ligação, do cenário menos preocupante ao mais arriscado para a sua segurança.
Pessoas reais
O cenário mais inofensivo é quando uma pessoa de fato ligou para você, mas o microfone dela está com algum problema. Talvez ela tenha silenciado a chamada por acidente com o ouvido ou o smartphone dela está conectado a um fone de ouvido Bluetooth, alto-falante ou sistema do carro que não está capturando sua voz. Falhas da operadora também podem silenciar o autor da chamada. A pessoa que faz a ligação pode nem perceber que há um problema; até onde ela sabe, ela está falando, mas ninguém pode ouvi-la. Nesses casos, você geralmente reconhece o número de telefone de quem liga.
Se a chamada for feita por um número desconhecido, ainda não há necessidade de entrar em pânico, embora a lista de pessoas que podem estar ligando fique muito maior.
Outra possibilidade legítima é um operador de uma central de atendimento que simplesmente não conseguiu colocar ou conectar o headset a tempo. Os sistemas das centrais de atendimento são projetados para discar números mais rápido do que os operadores conseguem se preparar. O sistema tentou encaminhar a chamada para um humano, mas nenhum operador estava disponível. É por isso que às vezes leva alguns segundos antes de você ouvir uma única palavra, podendo até mesmo ouvir tons de discagem como se fosse você quem estivesse fazendo a chamada.
Robô ou IA
Silêncio na linha é um sinal comum de chamadas feitas por robôs. Eles testam se um número de telefone está ativo e, se estiver, repassam a chamada para um humano, o que significa que um representante de vendas real (ou um golpista) ligará de volta para você nos próximos dias. Vale a pena notar que os golpistas não são os únicos que fazem chamadas com esse fim. Centrais de atendimento legítimas usam exatamente as mesmas ferramentas para reduzir a carga de trabalho dos seus operadores humanos.
Um agente de IA também pode estar por trás da chamada silenciosa. Para a pessoa que atende, não há diferença prática: a chamada parece idêntica a uma feita por um bot padrão. No entanto, a IA pode fazer mais do que apenas discar números automaticamente: ela pode analisar sua resposta e usar esses dados para decidir se seu número está ativo e se a chamada pode ser encaminhada a um humano para acompanhamento.
Golpistas
Agora vamos analisar a ameaça real. Talvez um dos tipos de chamadas silenciosas mais perigosas e desagradáveis sejam as feitas por um golpista. Uma ligação rápida e silenciosa como essa pode, na verdade, ser o primeiro passo de um golpe longo e elaborado, com histórias envolvendo empréstimos, órgãos do governo, outros golpistas e até autoridades policiais.
Empresas de cobrança também podem fazer esse tipo de ligação e permanecer em silêncio. Seu número pode acabar no radar deles se você, sua família ou alguém próximo tiver dívidas pendentes. Nesses casos, uma chamada silenciosa pode ser usada como uma tática de pressão psicológica.
Uma técnica semelhante é usada em casos de perseguição. Embora as ligações silenciosas, por si só, não causem danos diretos, elas podem ser usadas para gerar ansiedade, criar a sensação de que você está sendo observado o tempo todo e provocar desgaste emocional contínuo.
Por que as pessoas ligam e ficam em silêncio?
Quando você atende uma ligação, é provável que diga um rápido “Alô?” ou “Oi.” Isso é tudo o que a outra parte precisa para coletar uma grande quantidade de dados. Embora essas informações fossem difíceis de processar, o desenvolvimento da inteligência artificial tornou a tarefa muito mais fácil. Vejamos o que alguém pode descobrir sobre você ao falar apenas uma palavra:
Região, sotaque e localização. Os golpistas são sofisticados e astutos. Eles geralmente adaptam suas táticas a uma região, visando residentes de países específicos ou até mesmo de regiões próximas. Isso é muito relevante em países como a Índia ou a África do Sul, que têm 22 e 11 idiomas oficiais, respectivamente.
Idade e gênero aproximados. Embora uma pessoa possa facilmente confundir a voz de um adolescente com a de uma jovem ou estimar a idade de alguém de forma incorreta, a IA é muito mais eficiente em identificar nuances sutis da voz. Saber sua idade e gênero ajuda os golpistas a aprimorar suas estratégias para futuros ataques de engenharia social.
Horários em que você está disponível. Se você atende o telefone pela manhã, à tarde ou tarde da noite, os golpistas podem programar a ligação de acompanhamento para o período exato em que você tem maior probabilidade de atender.
Probabilidade de um ataque ser bem-sucedido. A IA pode avaliar automaticamente o valor potencial de um alvo. Por exemplo, se alguém atender rápido, falar com calma e não ignorar chamadas de números desconhecidos, é provável que essa pessoa receba uma prioridade mais alta para receber chamadas de acompanhamento por golpistas humanos.
A resposta da pergunta anterior “por que as pessoas ligam e ficam em silêncio” é coletar dados biométricos. Apenas alguns segundos de áudio gravado podem ajudar os cibercriminosos a criar um deepfake de voz. Embora uma ou duas palavras possam não produzir um clone convincente por si só, os golpistas juntam gravações de várias chamadas silenciosas para criar uma réplica confiável.
Essa tecnologia já está sendo usada em golpes do mundo real. Ao se passar por um parente, colega ou chefe, os golpistas podem solicitar que você envie dinheiro com urgência, compartilhe um código de autenticação de dois fatores para serviços governamentais ou conclua alguma outra solicitação que pareça não ter importância. Quanto mais realista for o deepfake, mais difícil é identificar o golpe, especialmente quando acompanhado de uma história convincente.
O que fazer ao receber uma chamada silenciosa?
Se você atender uma chamada, disser algumas palavras e desligar, não há necessidade de entrar em pânico. No entanto, essa breve interação pode confirmar aos golpistas que seu número está ativo e que você atenderá chamadas de números desconhecidos. Como resultado, seu número de telefone pode acabar em listas de alvos de futuras campanhas de spam ou golpes. Dito isso, é importante lembrar que uma única chamada silenciosa não representa uma ameaça imediata à segurança.
Ainda assim, é muito melhor se essas chamadas silenciosas nunca chegarem até você. É aí que os bloqueadores de chamadas como o WhoCalls podem salvar o dia. Trata-se de um aplicativo inteligente de identificação de chamadas que contém um enorme banco de dados de números conhecidos de fraudes e spam. É assim que funciona: você instala o aplicativo, concede as permissões necessárias e está tudo pronto! Além disso, é possível personalizar suas configurações de filtragem: bloquear chamadas recebidas de aplicativos de mensagens, escolher quais categorias de chamadas indesejadas rejeitar automaticamente e adaptar outras configurações de proteção às suas necessidades.
Veja adiante algumas dicas para ajudar você a manter a calma e evitar cair em golpes caso essas chamadas silenciosas acabem se tornando um problema:
Não atenda chamadas de números desconhecidos ou ocultos. Uma dica útil: se alguém realmente precisar entrar em contato com você, encontrará outra maneira de fazer isso ou continuará ligando do mesmo número em outros momentos. Os golpistas quase sempre utilizam números diferentes, enquanto os bots automatizados operam em uma programação rígida, como ligar todos os dias às 8h05.
Não se apresse para ligar de volta. Os golpistas geralmente contam com a curiosidade das vítimas para retornar ligações de números desconhecidos. Além disso, retornar a ligação pode resultar em cobranças se for um número de tarifa premium.
Não fale primeiro. Aguarde até que alguém do outro lado da linha inicie a conversa. Se você ouvir um ruído abafado ou silêncio, desligue e não perca seu tempo, pois é provável que se trate de um golpe.
Bloqueie números desconhecidos, mesmo após o término da chamada. Se você atendeu a ligação e percebeu que ela era suspeita, é melhor bloquear o número imediatamente. Para fazer isso, você pode usar aplicativos de bloqueio de chamadas ou os recursos presentes na maioria dos smartphones modernos.
Não compartilhe seu número em todos os lugares. Sites de phishing, páginas da Web que saem do ar de repente e brindes suspeitos geralmente existem apenas para coletar seus dados pessoais. Ao preencher formulários online, é preferível usar um número temporário ou secundário.
Obtenha um número de telefone secundário. Use números diferentes dependendo da ocasião. Compartilhe seu número principal com familiares, amigos e colegas de trabalho e use o número secundário para entregas, marketplaces online e para preencher formulários gerais na Web.
Scammers use AI deepfakes to impersonate OnlyFans creators, trick fans into sending money, then disappear after payment.
Criminals are building fake identities using AI-generated deepfakes of real OnlyFans creators, luring their followers with promises of live chats, and then disappearing after collecting payment. The scheme runs on social platforms that most people consider harmless, TikTok for discovery, Snapchat for the conversation, Cash App for the payment, and by the time the fan realizes something is wrong, the money is already gone and the account is blocked.
“This is a form of catfishing, in which an attacker impersonates someone online and engages in romantic or sexual interactions for ulterior motives. In this case, the scammers create fake accounts on platforms like TikTok, using material lifted from a real creator’s photos and given a synthetic voice.” reads the report published by MalwareBytes. “They’ll use that to nudge viewers into a direct message conversation on services like Snapchat.”
The choice of Cash App as the payment method isn’t accidental. It’s a peer-to-peer platform built for informal transfers between friends, not commerce — and transfers clear instantly. Once sent, the money is effectively gone.
“This kind of fraud has two victims: the fans who lose money to scams, and the creators. The latter lose income that they might have collected from fans, and also run the risk of retribution from disgruntled followers who think they’ve been taken advantage of.” continues the report.
One creator, Jessieanna Campbell, told USA Today she constantly receives messages from angry fans accusing her of taking their money and blocking them, for transactions she never made. Another creator interviewed by USA Today had fans show up at her home after the confusion, and now sometimes doesn’t feel safe leaving the house.
“The problem is that domestic laws only apply to domestic platforms. The stolen material largely sits on overseas hosts, making it difficult to control.” concludes the report. “Even if laws could be universally enforced, it might not matter. In three experiments conducted this year, researchers at the University of Bristol found that most participants relied on deepfake content even after being told it was fake.”
Federally, the Take It Down Act criminalizes non-consensual explicit content including AI-generated material and requires rapid platform takedowns. The EU’s AI Act requires disclosure when AI is used for image generation. Neither law has stopped the content from circulating because enforcement stops at national borders and the hosting doesn’t.
Spotting a deepfake is still possible if you know what to look for. USA Today described a TikTok video impersonating creator Elaina St. James, made by animating a still photo with a cloned voice, that showed distorted teeth and frozen eyebrows. Those artifacts are common in AI-generated video, especially when the source material is limited. Malwarebytes has published a practical guide to spotting deepfakes of any kind.
A similar pattern has been documented in romance scam contexts. Malwarebytes previously reported on Amazon and Apple impersonation scams using the same redirect-and-collect mechanics, where the victim is moved off a trusted platform into a direct channel before the ask. The underlying manipulation — establish familiarity, trigger urgency, request payment through an irreversible channel, is consistent across scam types regardless of which face or brand is being faked.
If a “creator” reaches out to you through a third-party platform and steers the conversation toward a direct payment for exclusive content, treat it as a red flag. Verify through the creator’s official, verified accounts before doing anything else. And if someone who looks like a creator you follow is asking for Cash App payment before delivering anything, assume you’re about to be blocked.
Researchers found that seven of nine tested Hugging Face image-editing tools produced sexualized alterations, highlighting gaps in model oversight, provenance, and enterprise vendor controls.
Those losses stem from 22,364 AI-related complaints. And these figures represent only the reported losses, which may well be the proverbial tip of the iceberg.
The main drivers behind the rise in AI-powered scams are voice cloning, deepfake images and videos, and AI‑generated scripts. These tools have supercharged classic fraud schemes such as romance scams, kidnapping and extortion calls, fake influencers, and government impersonation.
Michael Machtinger, deputy assistant director of the FBI Cyber Division, told the Wall Street Journal:
“AI-created fraudulent communications can look very official and very legitimate to even the most trained individuals.”
The FBI and financial institutions recommend verifying identities via official contact channels. One of their biggest concerns is government impersonation scams, which have evolved from crude IRS gift‑card phone calls into sophisticated, multi‑channel operations that combine spoofed caller ID, stolen agency logos, and AI‑generated audio and video of public officials.
This report, and others like it, shows how AI is being weaponized to automate research on victims, generate convincing scripts, and create highly believable deepfake personas at scale.
AI is also increasingly used in business email compromise (BEC), romance scams, and impersonation fraud. In BEC cases involving AI, losses have already reached tens of millions of dollars for businesses alone.
For a broader look at why AI is simultaneously fueling scams like these and becoming indispensable to defending against them, see my article AI: Threat, tool, or both?
It explains how both defenders and criminals use AI to find vulnerabilities, and why security vendors increasingly rely on AI to process vast amounts of telemetry, detect anomalies, and keep pace with threats that “no longer move at human speed.”
How to stay safe
Consumer protection agencies have documented a growing list of the ways scammers are using AI to try to rip people off. The main problem is that we can no longer take it at face value that the person we’re talking to is who they claim to be.
Government agencies and financial institutions recommend that you:
Be skeptical of urgent payment demands, especially those involving cryptocurrency or gift cards
Limit the amount of voice and video content you share publicly, as it can be reused by scammers
Report incidents quickly to your bank(s) and IC3.gov
Pro tip: Malwarebytes Scam Guard can help you determine whether a message is a scam and guide you through the next steps.
Something feel off? Check it before you click.
Malwarebytes Scam Guard helps you analyze suspicious links, texts, and screenshots instantly.
Content warning: This article contains descriptions of non-consensual sexual imagery.
Depending on which of his social media profiles you were looking at, Mark Resan was either a marketing lead at Google or working for a dental implant company, a human resources company and a business software firm – all at the same time.
Facebook photos showed Resan vacationing in Bali (left) and relaxing at luxury hotels in Dubai (right). Blurring by Bellingcat
But a Bellingcat investigation has found that the Hungarian national is the key figure behind, and the likely owner of, at least two deepfake porn websites – RefacePorn and DeepfakePorn – that until recently were selling paid subscriptions.
There is no question about the nature of these websites. RefacePorn’s landing page shows an explicit video of a woman performing a sexual act. As the video plays, her face is replaced with a variety of other women’s faces. The text above declares: “Face swap deepfake porn. Upload your face!”
Deepfake porn sites such as these, which use artificial intelligence to create sexually explicit images and videos – usually without the consent of those whose faces or bodies are featured – have proliferated at an alarming rate in recent years. The impact on victims has been described as “life-shattering”, with the mental health effects similar to those reported by victims of sexual assault.
While the technology to make these synthetic images is not new, the rise of mainstream AI image generator tools and “Nudify” apps has made it more widely available to people without deep technical expertise. Earlier this year, New Zealand MP Laura McClure held up an AI-generated nude of herself in parliament, describing how it took her less than five minutes to create after a quick Google search.
A 2024 study by the My Image My Choice campaign found that there was a 1,780 percent increase in sexually explicit deepfakes last year compared to 2019. Almost all (99 percent) of victims were women, according to a 2023 study by Security Hero.
Illustration for Bellingcat by Ann Kiernan
The creation of such images and videos is now illegal in a few countries, including the US and the UK, but legislation has not caught up in many others, and the owners of platforms that enable this content often face no repercussions. In May 2024, the EU passed a directive which mandates that member states – including Hungary, where Resan resides – criminalise the creation and distribution of non-consensual sexual deepfakes by June 2027.
Alexios Mantzarlis, co-founder of Indicator, a news site that focuses on digital deception, said his publication estimates that deepfake porn sites likely make millions of dollars a year.
“The incentive system will continue to exist until the tools become too toxic to handle for domain hosts and content delivery networks,” added Mantzarlis, who is also the director of the Security, Trust and Safety Initiative at Cornell Tech.
All Roads Lead to Resan
Bellingcat’s investigation into RefacePorn and DeepfakePorn – which spanned corporate registries, domain name registrations, payment redirect sites, website code and leaked data – led us back to Resan.
By simulating the purchase of subscriptions on these websites, Bellingcat was led through a series of redirects to a payments dashboard by Peerwallet, a payment processor that recorded more than US$331,000 in sales from July 2024 to August 2025 by Dorocron LLP. Dorocron is a Canadian-registered company whose main – if not sole – source of income appeared to be from paid subscriptions to these sites. The real amount is likely higher, as this was just one of several payment processors the websites have used.
Subscribe to the Bellingcat newsletter
Subscribe to our newsletter for first access to our published content and events that our staff and contributors are involved with, including interviews and training workshops.
Dorocron LLP did not respond to multiple requests for comment via email, and calls to the number listed on sites that had the company’s details in their legal information sections went unanswered.
Resan is the only person who appears to have been publicly associated with Dorocron LLP, and he is also the sole director of a UK-registered company, Facitic Ltd, that registered the domain of RefacePorn. Resan did not respond to multiple requests for comment sent via email over the past two weeks. Multiple emails and phone calls to Facitic Ltd also went unanswered.
However, days after we first reached out to Resan, his LinkedIn and X profiles were deleted, and his previously public Facebook profile was either deleted or made private. Both RefacePorn and DeepfakePorn also became inaccessible, displaying an error message that said “this site can’t be reached”.
Archives of RefacePorn and DeepfakePorn, which were previously available on the Internet Archive’s Wayback Machine, have also now been excluded from the archive. The Internet Archive told Bellingcat it processed exclusion requests submitted by someone with rights to both sites on Dec. 5.
Following the Money
Like other websites Bellingcat has investigated, RefacePorn’s ownership was hidden behind a network of website domains, fake websites used to redirect payments, and international business registries.
Using the tool DNSlytics, we examined the Google tag history on RefacePorn and found a tag that was also used on DeepfakePorn, as well as a website called facitic.com.
Google Analytics tags are small pieces of unique code that developers can place in the backend of a website to track its analytics. Each code is unique to a specific user, who can use the same tag across multiple websites.
Both RefacePorn and DeepfakePorn offer tiered subscription packages with similar names and prices based on the number of deepfakes that could be generated and the level of support.
When simulating a purchase of one of these packages – without actually completing payment – on DeepfakePorn, we received a link to make a payment hosted through the domain “remakerai.me”. Similarly, a mock purchase on RefacePorn pointed us to a payment link on “airemaker.me”. Bellingcat has observed the use of redirects, which can be used to obscure payments, by other deepfake porn sites. Many payment processors, including Paypal and Stripe, have restrictions on buying or selling sexually oriented online content.
Payment processors often block payments that come from
websites making deepfake pornography.
Using a redirect site hides the original site from the
payment processor, making it harder to block.
Despite this, payment processors sometimes manage to
block the redirect site.
But If one redirect site is blocked, the site owner can
quickly switch to another redirect site that isn’t
blocked.
Graphic: Galen Reich
The redirected payment links hosted on airemaker.me and remakerai.me offered several payment options including Paypal, credit cards and cryptocurrencies. Bellingcat selected the credit card option, and in both cases was emailed a link to complete the purchase on a payment platform called Peerwallet. This email included a link to the seller’s profile, Dorocron LLP.
This profile showed the funds received by the seller, which totalled more than $331,000 as of August 2025. This income was related to 16,264 sales. According to this dashboard, Dorocron LLP had been a member of Peerwallet since July 22, 2024, meaning these sales all occurred over the past year.
Screengrab of Peerwallet profile for Dorocron LLP, showing about US$331,000 in funds received for sales
RefacePorn has been active since at least May 2022, according to promotional posts by an Instagram account with the username “Dorocron2323” and the account name “Hassler Mark”. Social media accounts for RefacePorn were also created on X and Facebook in May 2022.
Screengrab of an Instagram post from May 2022 promoting RefacePorn’s website, which is now down. Blurring by Bellingcat
While the transactions on Peerwallet were not broken down by domain, two were the payment redirect sites for the deepfake porn sites we investigated. Bellingcat’s review of the 21 “approved domains” listed on this profile found no evidence that payments were ever accepted through the other sites.
Short-lived, “disposable” domains are known to be used by bad actors to evade detection, presenting a moving target for payment processors and authorities. As of publication, both airemaker.me and remakerai.me are no longer accessible. But in the course of the investigation, we observed RefacePorn and DeepfakePorn’s payment links redirecting to other third-party sites, before the sites went offline.
The Peerwallet profile showed transactions by users, as well as 21 approved domains including those redirecting payments for RefacePorn (refaceporn.com) and DeepfakePorn (deepfakeporn.app)
Of the 21 domains on Dorocron LLP’s Peerwallet profile, only two were still accessible as of the end of November, with the rest either down due to expired domains or server issues, displaying generic domain parking pages, or requiring a login to view. Though almost all of the sites had their registration information redacted, Resan was listed as the most recent registrant for one of the expired domains.
The two sites still accessible listed a variety of products, including eBooks and digital products. Both had almost identical products and templates, and listed Dorocron LLP under their company information in their footers.
Bellingcat tried to check out items on each of the sites, and in both cases was prompted to log in. It was, however, impossible to register an account, and when we tried with an active email address we were redirected to a login page saying that the email address was “unknown”.
Archived screengrabs of some of the sites that now have expired domains or require a login to view showed that many of them followed the same format, selling eBooks and video courses with “resell rights”.
Peerwallet told Bellingcat in September that Dorocron LLP was “not approved” to sell deepfake porn, and that it was looking into the issue. However, when Bellingcat asked for an update in November, Peerwallet appeared to have closed down. Emails to the payment processor’s founder have also gone unanswered.
The Man Behind the Screen
Dorocron LLP was registered in British Columbia, Canada in March 2022. We were unable to verify if Resan’s name was on the corporate records as information on company owners or directors in British Columbia is restricted to law enforcement and other officials.
However, Resan’s name has been used to register at least 13 sites alongside an email bearing Dorocron’s name from as far back as 2013, nine years before Dorocron was registered in Canada. The earliest domain registration, from 2013, included the name of a now-dissolved UK-registered company called “Webnaser LTD”, whose registration documents also cite Resan as the sole director.
A leak found on data breach site Intelx.io shows that an almost identical password (with different capitalisation of some letters) was used to log into this “dorocron” Gmail account and a Netflix account associated with Resan’s personal email address. This password was also used to log into web domain registry GoDaddy using RefacePorn’s support email address.
Leaked passwords on Intelx.io revealed another link between Resan and DeepfakePorn: an email with the username “resanmark” was used to log into DeepfakePorn’s website, with a password containing his birth year. In all, we found four unique passwords that were reused between Resan’s personal emails, the Dorocron emails, and a support email for RefacePorn. These four passwords include either Resan’s name or the date or year of his birth.
Resan also posted two job listings from his now-deleted LinkedIn account about a year ago, for a full-stack web developer and a WordPress developer at Dorocron LLP. In the web developer listing, he described the company as “developing and applying revolutionary AI technologies” and said the job would have “high wages”. We could not find any other individual with a public association to Dorocron LLP on LinkedIn or elsewhere.
Support Bellingcat
Your donations directly contribute to our ability to publish groundbreaking investigations and uncover wrongdoing around the world.
Aside from his links to Dorocron LLP, Resan is also the sole director and person with significant control of Facitic Ltd, a UK-registered company which was listed as the registrant for RefacePorn.
Using DomainTools, we were able to see the historical registrant information in a WHOIS lookup of the site’s domain registration. When we checked this in August 2025, we were able to see that, as of June 2025, Facitic Ltd was the registered owner of RefacePorn. This information was later redacted – as it is for other sites linked to Resan such as DeepfakePorn.
ICANN, which regulates websites, requires domain name providers to verify the accuracy of their customers’ details, including the registrant's name and contact details. Such details are publicly visible by default, but can be anonymised using paid privacy services.
The UK registration for Facitic Ltd lists Resan’s country of residence as Dubai, while the registration for another UK company he registered – which was also listed as the owner of some of the now-expired approved domains on Dorocron LLP’s Peerwallet profile – states that he resides in Cyprus. Meanwhile, Resan’s social media accounts stated that he lives in Hungary. On Peerwallet’s dashboard, the primary user of Dorocron is listed as being based in Hungary.
It is unclear if Resan actually holds positions in any of the six companies he listed himself as working at on his Facebook and LinkedIn profiles. Bellingcat has reached out to these companies to check, but has not received any replies as of publication.
Some of the connections Bellingcat found between RefacePorn and Mark Resan:
Graphic: Galen Reich
On Nov. 10, 2025, a few weeks before we contacted him, Resan applied for Facitic Ltd to be struck off the UK companies register. Based on Resan’s filings, Facitic Ltd was incorporated with an initial capital of £100 in January 2024, and there has been no recorded change in its accounts since.
It is unclear what triggered Resan to file to dissolve the company, and he did not respond to Bellingcat’s query about this.
Small Sites, Big Harm
The websites linked to Resan are not among the largest in the deepfake porn industry. A similar but much larger site that Bellingcat has investigated, MrDeepFakes, received millions of visits each month. Bellingcat and its partners Tjekdet, Politiken and CBC exposed the site’s key administrator David Do in May, with MrDeepFakes going offline after we reached out to Do for comment.
In comparison, RefacePorn and DeepfakePorn received about 91,000 and 154,000 visits in October, according to digital marketing platform SemRush. But their smaller size does not mean they can’t cause significant harm.
Mantzarlis, of the news site Indicator, said there were “smaller players” taking bigger risks around regulation, such as “Crush AI”, a group of Chinese-owned apps that bypassed Meta’s moderation rules to run 25,000 ads on Facebook and Instagram before the social media giant sued them.
“These smaller players are often the ones that are more actively trying to stand out on social media to catch up with the bigger ones,” Mantzarlis said.
In the course of our investigation, we ran tests using the free features on RefacePorn to determine if there were any restrictions on images that could be uploaded on the website.
Without actually generating the content, we uploaded AI-generated images of adult women and underage girls. Unlike on other websites we have tested, which have added the bare minimum of checks to prevent uploading images depicting children, there was no restriction or evidence of age-related safeguards on RefacePorn.
While there aren’t laws in Hungary explicitly prohibiting deepfake porn, the possession, creation and distribution of sexually explicit images of minors is illegal.
“As the more established websites come under sustained regulatory pressure and others get litigated into oblivion, the minnows are ready to try and capture market share,” Mantzarlis said.
And while some sites such as RefacePorn and DeepfakePorn may fold in the face of public scrutiny, others continue to operate, unchecked and easily accessible, online.
“These websites are eminently replaceable and there's no reason to believe that there is any form of ‘brand loyalty’,” Mantzarlis said. “Perpetrators are going to search for ‘nudify’ or click on an ad and go to whatever tool does the job.”
Melissa Zhu contributed to this report.
Bellingcat is a non-profit and the ability to carry out our work is dependent on the kind support of individual donors. If you would like to support our work, you can do so here. You can also subscribe to our Patreon channel here. Subscribe to our Newsletter and follow us on Bluesky here and Mastodon here.
Neste ano, o Brasil sofreu mais de 315 bilhões de tentativas de ataque cibernético apenas no primeiro semestre — número que representa cerca de 84% de todo o tráfego malicioso registrado na América Latina no período, de acordo com um levantamento da Fortinet divulgado em agosto passado. Com adversários fazendo uso crescente de IA para conduzir campanhas de phishing, ataques DDoS e exploração de vulnerabilidades em ambientes híbridos e multicloud, líderes de segurança corporativa se veem diante de uma encruzilhada: manter o SOC (Security Operations Center) da forma como está ou apostar em novas tecnologias e formas de trabalho que respondam à nova realidade.
Grupos criminosos adotam táticas avançadas e combinam técnicas de Advanced Persistent Threat (APT) com ferramentas de IA para driblar as defesas tradicionais. Golpes que envolvem deepfakes ou phishing automatizado por IA tornam cada vez mais difícil distinguir o legítimo do malicioso. Ao mesmo tempo, a superfície de ataque das organizações se expandiu dramaticamente. Com ambientes de TI híbridos e multicloud, centenas de novos serviços e integrações são adicionados constantemente aos ecossistemas corporativos, abrindo brechas que muitas vezes são difíceis de monitorar.
O problema é que grande parte dos SOCs atuais foi concebida para um contexto em que as ameaças eram baseadas em assinatura e o volume de eventos era controlável. Hoje, porém, a multiplicação de fontes de telemetria, tais como endpoints, nuvens, APIs, identidades, dispositivos IoT e aplicações SaaS, faz com que o volume de logs cresça de forma exponencial.
Não se trata apenas de mais dados, mas de dados de naturezas distintas, com diferentes formatos, níveis de granularidade e relevância operacional. Ferramentas modernas de detecção e monitoramento, como EDRs, XDRs e soluções de observabilidade, coletam informações em tempo real e geram fluxos contínuos de telemetria que precisam ser correlacionados com ameaças conhecidas e comportamentos anômalos. Sem uma arquitetura de dados e automação adequadas, esse ecossistema torna-se difícil de orquestrar – e o SOC passa a lidar menos com ruído e mais com complexidade analítica. O desafio, portanto, deixou de ser filtrar falsos positivos e passou a ser transformar grandes volumes de logs em inteligência acionável com velocidade e contexto.
E, quando os adversários passam a empregar IA para criar malwares praticamente indetectáveis, um SOC calcado apenas em esforço humano não consegue escalar na mesma proporção do risco. O resultado é um descompasso perigoso entre a capacidade defensiva e a velocidade com que as ameaças modernas atuam, evidenciando que manter o status quo não é mais sustentável.
Uma das principais transformações em curso é a adoção do modelo de SOC as a Service, que redefine a forma como as empresas estruturam sua defesa cibernética. Diferente do modelo híbrido ou totalmente interno, o SOCaaS oferece monitoramento, detecção e resposta a incidentes 24×7 por meio de uma plataforma escalável e baseada em nuvem, administrada por especialistas em cibersegurança.
Esse formato elimina a necessidade de manter infraestrutura local pesada e reduz o tempo de implantação, ao mesmo tempo em que garante acesso contínuo a tecnologias e analistas altamente especializados.
Ao integrar telemetria proveniente de múltiplas camadas, o SOC as a Service consolida os eventos em um único datalake de análise, aplicando correlação e contextualização automatizadas com apoio de SOAR e machine learning. Assim, os alertas deixam de ser tratados de forma isolada e passam a compor narrativas completas de ataque, permitindo uma visão tática e antecipada das ameaças.
Essa automação nativa reduz drasticamente o tempo médio de detecção (MTTD) e o tempo médio de resposta (MTTR), pontos críticos para conter ataques modernos que podem se propagar em minutos.
Outro benefício do modelo é a atualização contínua da inteligência de ameaças. Fornecedores de SOCaaS normalmente operam com bases globais de threat intelligence, alimentadas por fontes de ciberinteligência regionais e internacionais.
Essa atualização constante amplia a visibilidade sobre novas campanhas maliciosas, técnicas de exploração e indicadores de comprometimento (IoCs), garantindo que o ambiente corporativo permaneça protegido mesmo diante de vetores inéditos. Ao mesmo tempo, as plataformas de SOCaaS modernas integram recursos de análise comportamental (UEBA) e aprendizado contínuo, permitindo identificar padrões anômalos e prevenir movimentos laterais antes que evoluam para incidentes graves.
Mais do que uma modernização tecnológica, a adoção de modelos de SOC as a Service representa um novo paradigma de defesa cibernética. O CISO que ainda vê o SOC apenas como um centro de monitoramento precisa agora encará-lo como um núcleo de inteligência e antecipação, sustentado por automação, correlação de dados e aprendizado de máquina.
Entrar
