A brand new pattern has gone viral on social media, the place customers share a private photograph and ask Synthetic Intelligence (AI) instruments to create a caricature or illustration based mostly on their life, their job, and “every part the AI instrument is aware of” about them. The end result displaying animated variations of the individual on the workplace, with their household, or representing their career has turn out to be in style content material on Instagram, LinkedIn, and even X.
Whereas the pattern could appear inventive and entertaining, cybersecurity specialists warn that this apply can expose private info and allow the creation of personalised, large-scale fraudulent messages, a really actual menace right this moment.
Based on Adrian Hia, Managing Director for Asia Pacific at Kaspersky, the sort of request doesn’t work like a easy visible filter. “To realize extra correct photos, folks permit AI instruments to entry all the knowledge related to their profiles with out restrictions, because the instruction itself is embedded within the command ‘create a caricature about me and my job based mostly on every part you already know about me’. Along with the reference photograph, additional information reminiscent of firm title, company logos, job title, metropolis, every day routines, hobbies, and different household particulars are sometimes included and used to create the pattern,” he stated.
What sort of private information is uncovered?
“That is the query most individuals don’t suppose to ask—as a result of the output seems to be like a cartoon, not a file. However pause and take into account what goes in. A photograph alone reveals your face, approximate age, ethnicity, and sometimes a background that hints at your private home or office. If you happen to add a immediate like ‘every part you already know about me,’ and the AI attracts from no matter you’ve gotten shared in that session or linked account—your title, job title, employer, household references, metropolis, pursuits, even your communication model. What comes out the opposite aspect isn’t just a caricature. It’s a synthesised private profile, assembled in seconds, that you just willingly handed over,” Dr Sanjay Katkar, Joint Managing Director at Fast Heal Applied sciences Restricted, informed indianexpress.com.
How can sharing info make scams extra convincing?
“Combining an individual’s {photograph} with contextual particulars considerably will increase the danger of phishing or impersonation scams as a result of it permits scammers to create extremely plausible identities. If scammers receive an individual’s {photograph}, they will use it to create faux social media or skilled profiles throughout totally different platforms. The picture will also be used to generate voice or video deepfakes, or to clone firm or skilled profile pages the place the sufferer’s photograph is used because the show picture, making the profile seem genuine,” says cybersecurity professional Anurag Mathur.
He provides, “If job or office particulars can be found together with the {photograph}, scammers can exploit this info to craft convincing enterprise emails or messages. By figuring out the individual’s job title, place within the organisational hierarchy, and the corporate they work for, fraudsters can impersonate the person and ship faux directions or requests to colleagues or staff inside the organisation.”
“Equally, if details about relations is understood, scammers might try emotional manipulation. They’ll ship pressing messages or calls to kin claiming {that a} member of the family has met with an accident or is in bother, and demand instant monetary assist or ask them to hurry to a selected location,” Mathur notes.
Story continues beneath this advert
Hia mentions, “Every of those information factors is a key piece in constructing an in depth digital profile. By combining picture, textual content, and context, habits, relationships, steadily visited locations, {and professional} tasks are revealed—info that cybercriminals can exploit to craft extra refined scams. Because of this, a fraud try that mentions the place somebody works, their job title, or perhaps a member of the family turns into way more convincing and will increase the probability that the sufferer will belief it and share delicate info or cash.”
How helpful is an AI-generated private profile to scammers?
“Extraordinarily helpful. Cybercriminals normally spend appreciable time conducting open-source intelligence gathering (OSINT), piecing collectively fragments of an individual’s digital footprint from platforms like LinkedIn, Instagram, information articles, or information breaches. Viral AI caricature tendencies can successfully try this work for them in a single consolidated output. As a substitute of scraping info from a number of sources, scammers get a ready-made profile containing a face, persona cues, skilled context, and private particulars,” Katkar opines.
Amit Relan, CEO and co-founder of mFilterIt, says, “What makes this pattern uniquely harmful as a fraud channel is the aggregation impact. No single piece of knowledge a consumer shares is dangerous in isolation—however the AI synthesises all of it right into a consolidated identification profile: your office, your anxieties, your relationships, your habits. That synthesis is strictly what a fraudster must craft a phishing assault you’ll imagine, an impersonation you’ll not query. A rip-off tailor-made exactly to your blind spots.”
Hia says this threat is especially acute within the APAC area. Regardless of a excessive AI adoption fee, with 78 per cent of execs utilizing AI weekly (surpassing the worldwide common of 72 per cent), many customers nonetheless battle with primary technical literacy, leaving them weak to social engineering and phishing.
Story continues beneath this advert
As well as, when interacting with these platforms, customers should not solely sharing the ultimate picture. Relying on the service and its privateness insurance policies, the unique photograph, the consumer’s textual content or directions, utilization historical past, and sure technical information reminiscent of IP handle, gadget, or interplay patterns may be saved. A part of this info could also be retained to function the service, enhance efficiency, or prepare AI fashions, that means the content material doesn’t essentially disappear after the caricature is generated and should stay longer than customers anticipate.
“This viral pattern of caricature creation of our lives might look like innocent enjoyable, however it’s successfully a voluntary briefing for cybercriminals,” says Hia. “In a area the place AI adoption is main the world however technical literacy continues to be catching up, these digital portraits have gotten harmful maps,” provides Hia.
Right here’s what you are able to do
Whereas these instruments generally is a enjoyable strategy to experiment with digital creativity, consultants indianexpress.com spoke to suggest adopting extra cautious habits:
📍Keep away from coming into identifiable information in prompts reminiscent of your full title, job title, firm, metropolis, handle, schedules, or every day routines, even when it looks like it is just for personalising the picture.
Story continues beneath this advert
📍 Don’t add images that present logos, credentials, paperwork, licence plates, screens, constructing façades, or another aspect that would assist determine your location or hyperlink you to an organisation.
📍 Keep away from importing images that clearly present your face together with identifiable particulars reminiscent of kids, ID playing cards, or office badges.
📍 Don’t share photos or details about minors, and keep away from revealing household particulars that might be used to impersonate kin or create emotional scams.
📍 Don’t use prompts reminiscent of “every part you already know about me.” As a substitute, specify: “Create utilizing solely the main points on this message and don’t use previous chats or historical past.”
Story continues beneath this advert
📍 Crop or take away apparent identifiers in images earlier than importing them, reminiscent of avenue indicators, certificates, constructing names, or society title boards.
📍 Keep away from together with exact geographic info in prompts, for instance, “me as a Bangalore engineer working at XYZ Tech Park.”
📍 Don’t add images of different folks with out their consent.
📍 Disable chat historical past or data-saving options if the platform supplies that possibility and you don’t want your information reused.
Story continues beneath this advert
📍 Assessment the platform’s privateness coverage and permissions earlier than utilizing it, particularly relating to how lengthy content material is saved and whether or not it’s used for mannequin coaching or service enhancements.
📍 Think about using privateness modes reminiscent of ‘Momentary Chat’, the place prompts, images, and outputs should not retained past a restricted time or used for coaching AI fashions.
📍 If you wish to strive the pattern, use an older or edited picture as a substitute of a high-resolution present {photograph} and test whether or not the platform lets you request deletion of your information afterwards.
📍 As an added layer of safety, safety instruments may also help cut back dangers from malicious hyperlinks, harmful downloads, and phishing makes an attempt related to viral tendencies.
Story continues beneath this advert
📍 Most significantly, earlier than taking part in any viral AI pattern, ask your self: What private info am I truly sharing? Keep away from importing high-resolution images or identifiable private particulars you probably have not reviewed the platform’s information retention insurance policies.
The Protected Facet
Because the world evolves, the digital panorama does too, bringing new alternatives—and new dangers. Scammers have gotten extra refined, exploiting vulnerabilities to their benefit. In our particular characteristic sequence, we delve into the newest cybercrime tendencies and supply sensible suggestions that can assist you keep knowledgeable, safe, and vigilant on-line.