Digital Natives Are Not Fraud-Proof: How AI Turned Fluency Into Exposure

The guidance arrived with the timetable. In the last week of August 2026, as American students loaded cars and signed tenancy agreements, JPMorganChase pushed a piece of consumer advice out through local news partners with a claim that sat oddly against the usual back-to-school fare about textbook scams and dodgy rental listings. Today's students, the bank wrote, are true digital natives who have grown up online and who turn to digital platforms for everything from apartment hunting and job searches to shopping and banking. Then came the turn: that familiarity, however, can create a false sense of security, making younger adults just as much of a target for fraud and scams as older individuals.

The guidance is widely dated to the first week of September, when syndicated versions were still landing as students returned. The earliest version that can be verified ran on 21 August 2026, and variants continued appearing into early September, one of which quoted Diedra Porché, JPMorganChase's head of community and business development, warning students that scammers try to lure people into a false sense of urgency and that they should not rely on pictures or videos sent to them, because those could be completely fake. The dating matters less than the argument, which is genuinely unusual coming from a retail bank. Banks normally tell older customers to be careful. This one told the cohort that has never known a world without broadband that its lifelong fluency is not a defence, and attached a number: according to the FBI's Internet Crime Report, Americans aged 20 to 29 reported more than 500 million dollars in internet-crime losses in 2025.

That number is accurate, and it is also one of the most misleading figures in American fraud statistics, for reasons that have nothing to do with the FBI's arithmetic and everything to do with what the number is capable of counting.

What the Age Table Actually Says

The FBI's Internet Crime Complaint Centre published its 2025 annual report in April 2026, recording 1,008,597 complaints and $20.877 billion in reported losses, an average of $20,699 per complaint. Its age table breaks down as follows. Under 20: 31,254 complaints, $67.1 million. Ages 20 to 29: 112,069 complaints, $563.1 million. Ages 30 to 39: 153,293 complaints, $1.7 billion. Ages 40 to 49: 167,066 complaints, $2.957 billion. Ages 50 to 59: 124,820 complaints, $3.7 billion. Ages 60 and over: 201,266 complaints, $7.7 billion.

So the bank's headline is correct on its face. It is what happens when you set the 2025 figures against 2024 that the shape of the thing appears. In 2024, the 20 to 29 bracket filed 71,399 complaints and reported $540.1 million in losses. In 2025 the complaint count rose by 57 per cent while the reported losses rose by around 4 per cent. Per complaint, the average loss in that bracket fell from roughly $7,600 to roughly $5,000. Over the same period, losses reported by the over-sixties climbed from $4.8 billion to $7.7 billion, an average of roughly $38,500 per complaint.

Two very different crimes are hiding under one headline. Among the young, incidents are multiplying and individual losses are shrinking. Among the old, incidents are rising modestly and individual losses are catastrophic. The under-twenties illustrate the first pattern most starkly of all: complaints rose 74 per cent year on year while reported losses tripled from $22.5 million to $67.1 million, off a base so small that any movement looks dramatic.

The Federal Trade Commission's Consumer Sentinel data, gathered by a different route with a different population, says the same thing in cleaner form. In its 2024 data book, published in March 2025, 44 per cent of fraud reports filed by people aged 20 to 29 involved an actual loss of money, against 24 per cent of reports from people aged 70 to 79. But the median loss for the twenty-somethings was $417, against $1,000 for the seventies and $1,650 for the over-eighties. Younger adults get hit far more often and lose far less each time. This is the single most robust finding in the entire literature on age and fraud, it has held for years, and it is the finding that both the “young people are savvy” story and the “young people are just as vulnerable” story manage to obscure.

Where the 563 Million Actually Went

Decompose the 20 to 29 figure and the composition is the argument. Complaints involving cryptocurrency from that age bracket accounted for $288.9 million of losses across 18,107 complaints, which is more than half of everything the cohort reported losing. Cryptocurrency investment fraud specifically accounted for 4,627 complaints and $117.4 million.

The most common crime type in the bracket by volume was not investment at all. It was extortion, at 26,963 complaints, followed by phishing and spoofing at 19,765. The IC3 received more than 75,000 sextortion submissions in total during 2025 and referred more than 5,700 involving minors to the National Centre for Missing and Exploited Children. Sextortion complaints from the 20 to 29 bracket numbered 22,061 and carried $7.28 million in adjusted losses, an average of a few hundred dollars each. The 60-plus bracket filed 6,121 sextortion complaints and lost $14.9 million, roughly twice as much money from a little over a quarter as many reports.

Employment fraud accounted for another 4,555 complaints. The FTC has been tracking the mechanism behind much of that: gamified task scams, in which a target is recruited to perform repetitive online tasks, shown accumulating earnings, and then required to deposit their own money to unlock a withdrawal. Reports of these went from effectively nothing in 2020 to about 20,000 in the first half of 2024 alone. In April 2026 the FTC reported that $2.1 billion was lost in 2025 to scams that began on social media, roughly eight times the 2020 figure, and that the 18 to 29 bracket had the highest share of any age group of fraud loss reports originating there.

Add it up and the portrait of the defrauded twenty-year-old is not a person taking a cloned-voice emergency call. It is a person who bought a token, answered a direct message, applied for a job, or sent a rental deposit. The crimes concentrated on this cohort run through exactly the platforms in which its members are most practised. Fluency did not protect them. Fluency is what put them in the room.

A Term Coined Without Evidence

It is worth being precise about the phrase everyone is using. “Digital native” was popularised by the education writer Marc Prensky in 2001 as a rhetorical device about pedagogy, not as an empirical finding about capability. The research literature spent the next two decades failing to support it and eventually turning hostile.

The most cited demolition is by Paul Kirschner and Pedro De Bruyckere, whose paper “The myths of the digital native and the multitasker” appeared in Teaching and Teacher Education in October 2017. Their conclusion was that the digital native does not exist as a distinguishable category of competence, and that treating young people as though they possess innate technological sophistication actively harms them, because it licenses institutions to skip the teaching. Nature ran an editorial under the flat headline “The digital native is a myth” in July 2017, making the related point that ease with devices and understanding of the systems behind them are unrelated capacities.

More recent work has taken the question directly to fraud. Dewan Mehrab Ashrafi, Mahnoor Sattar and Rafayat Jalal, writing in the Journal of Financial Crime in late 2025, surveyed 385 self-identified digital natives on phishing susceptibility and found that psychological and psychographic factors dominated socio-economic ones, with risk tolerance the strongest single predictor and anxiety increasing vulnerability. Being young and online did not confer resistance. Personality did some of the work that competence was assumed to do.

Then there is the confidence problem, which is measurable. In February 2025 the biometric firm iProov tested 2,000 consumers in the United Kingdom and United States on genuine and synthetic media. Only 0.1 per cent correctly classified every item. Participants were 36 per cent less likely to spot a synthetic video than a synthetic image. And the adults aged 18 to 34, who performed no better than anyone else, were the most confident that they had performed well. That gap between measured accuracy and self-assessment is the blind spot in its purest laboratory form. It is not ignorance. It is calibration failure, and calibration failure is worse, because it suppresses the impulse to check.

The Tells That Stopped Telling

The defences this cohort genuinely does possess are real, hard-won and specific. They are pattern-matching defences, built through years of exposure to a particular industrial process. Misspellings. Mangled grammar. A URL with one letter wrong. A greeting that addresses you as “Dear Customer”. A story about an inheritance in a country the sender cannot spell. These heuristics worked because they were not really detecting fraud. They were detecting the presence of a human being writing quickly in a second language for very little money.

Generative text removed that signal in about eighteen months. The clearest experimental evidence comes from Fred Heiding, Simon Lermen, Andrew Kao, Bruce Schneier and Arun Vishwanath, who ran phishing emails past human subjects and published the results at the end of 2024, with a journal version following in Expert Systems with Applications. A control group of arbitrary phishing emails achieved a 12 per cent click-through rate. Emails written by human experts achieved 54 per cent. Emails generated end to end by an AI system, with no human involvement in reconnaissance or drafting, also achieved 54 per cent. Adding a human to the loop got them to 56 per cent. The authors estimated the automated pipeline cut campaign costs by more than 95 per cent.

Read that finding through the eyes of someone who has spent fifteen years learning to spot bad English. The skill they trained is not merely obsolete; its inverse is now dangerous. A message that is fluent, correctly punctuated, contextually specific and addressed to you by name used to be weak evidence of legitimacy. It is now evidence of nothing at all, and the reflex that treats polish as reassurance is a liability that older users, who never fully trusted the medium in the first place, are less likely to carry.

Three Seconds, Stated Accurately

The claim that a voice can be cloned from a three-second sample circulates constantly and deserves careful handling, because it is nearly true in a way that overstatement can only damage.

The number originates in research rather than marketing. Microsoft researchers published VALL-E in January 2023, describing a neural codec language model trained on 60,000 hours of English speech that could synthesise personalised speech from a three-second enrolled recording of a speaker it had never encountered. That is a laboratory result under controlled conditions, and it describes a capability rather than a product. In the same year McAfee's security researchers tested consumer voice-cloning tools and reported that three to four seconds of audio produced roughly an 85 per cent voice match using a free tool, and that training on a small number of additional files pushed the match to around 95 per cent.

The accurate statement, then, is not that any voice can be perfectly reproduced from three seconds. It is that three seconds is enough for an approximation good enough to survive a compressed, emotionally charged telephone call, and that more audio makes it better. This distinction matters because the accuracy threshold required for fraud is far below the threshold required for, say, an audiobook. A frightened person on a bad mobile connection is not conducting a forensic comparison. Three seconds is roughly the length of a voicemail greeting, an Instagram story, or the audio bed under a video posted to a public profile. For the demographic that has documented its own voice more comprehensively than any in history, the raw material was volunteered years ago and cannot be recalled.

When the Caller Costs Nothing

The constraint that has always limited telephone fraud is labour. Somebody has to be on the line, awake, in a working accent, able to improvise when the target asks an unexpected question. That constraint set the economics: it made sense to call people with money, because a caller's hour spent on a low-value target was an hour wasted.

Sanket Badhe of Rutgers University published a paper in 2025, presented at the Conference on Applied Machine Learning for Information Security, describing a system called ScamAgent. It is an autonomous multi-turn agent that maintains dialogue memory, adopts persistent personas, adapts to a target's responses and pursues a sequence of sub-goals: establish credibility, build trust, elicit sensitive information, avoid tripping the model's own safety filters. Badhe tested it across scenarios including medical insurance verification, prize and lottery fraud, government benefit enrolment and impersonation.

Two findings deserve attention. The first concerns guardrails. Asked directly, in a single prompt, to produce a scam script, frontier models refused between 84 and 100 per cent of the time. Routed through the agent framework, where the malicious objective was decomposed across turns and carried in context rather than stated, refusal rates fell to between 17 and 32 per cent. Safety training that works on a sentence does not necessarily work on a conversation. The second concerns quality. Five independent human raters scored ScamAgent's dialogues against transcripts of real scam calls on a five-point scale, giving the synthetic conversations 3.4 for plausibility against 3.6 for the genuine article, and 3.6 for persuasiveness against 3.9. The machine was very slightly worse than professional criminals, and it was converted into audio by off-the-shelf text-to-speech.

Slightly worse, at effectively zero marginal cost, is the whole story. When the caller costs nothing, the rational target set expands to include everyone, and it becomes economically sensible to make a hundred thousand calls chasing a $417 median loss. The cohort that was previously beneath the notice of a labour-limited industry has just been brought inside its addressable market. That, rather than any change in young people's judgement, is the most plausible explanation for a 57 per cent jump in complaints from twenty-somethings in a single year.

Inoculation Rather Than Information

If information were the answer, China would have solved this. Cyber fraud accounts for over half of criminal cases there, and anti-fraud education is close to universal. Which is what makes the study published in January 2026 by Xiangzhe Yuan, Jiajun Wang, Huanchen Wang, Qian Wan and Siying Hu so instructive, and why it should be read more carefully than its summaries suggest.

The system, whose name appears on the arXiv listing as ImmuniFraug and which circulates elsewhere as ImmuniFraud, is a large language model driven simulator that puts undergraduates inside multimodal fraud scenarios across ten categories, complete with voice, avatars, branching choices, escalating pressure and a debrief grounded in protection motivation theory. The trial ran with 846 Chinese undergraduates, 391 in the experimental condition and 455 in the control. Ninety-two per cent of participants reported prior exposure to anti-fraud training, so the study was measuring what simulation adds on top of near-saturation information.

It added something, and the something was small. The improvement in fraud awareness scores was statistically significant at p equals 0.026, with a Cohen's d of 0.151, which the authors themselves describe as a modest practical effect. The outcome measured was a fraud awareness scale, not behaviour under a live attack, and the strongest results were in self-reported constructs: perceived self-efficacy, behavioural intention, and a favourable comparison against traditional materials. Participants said the realism, the time pressure and the emotional manipulation were what made it work.

That is a genuine finding, and it points somewhere uncomfortable. If being told about fraud produces almost nothing in a population already told everything, and being deceived in a simulation produces a small but real increment, then effective education means institutions systematically deceiving the people in their care, repeatedly, with enough fidelity to induce actual panic, and then debriefing them. Corporate security has done a version of this with phishing simulations for a decade and it remains ethically contested, partly because the intervention works by damaging trust in the institution running it. A university that clones a student's mother's voice to teach them a lesson has taught them something. It has not necessarily taught them the thing it intended.

The Family's Designated Verifier

The most quietly devastating of the three papers is the one that is not about young people at all. In August 2025 Karina LaRubbio, Alyssa Lanter, Seihyun Lee, Mahima Ramesh and Diana Freed published a study of deepfake impersonation scams based on focus groups with 37 adults aged 65 and over, conducted in the north-eastern United States in early 2025.

The finding that matters is structural. Every protective practice the participants described depended, to some degree, on a younger family member. One recounted advice they used constantly: if it is that important, they will leave a message. Another described a plan to interrogate the caller and then telephone a relative independently. A third described the successful defence of an attempted grandparent scam through shared knowledge: asked who was calling, they replied that they had three grandsons, so which one was he, and the line went dead.

Two things follow. First, the young person is frequently the family's authentication infrastructure: the person who sets up the code word, advises against answering unknown numbers, and gets rung to confirm that the emergency is real. Second, as the authors note through what they call a dual victimisation framework, the same young person's likeness and voice are the raw material from which the impersonation was built. They are simultaneously the target of the theft and the last line of defence against it.

That is what makes overconfidence a systems failure rather than a personal one. A family's verification procedure has a single point of failure, and it is located in the member most certain that they cannot be deceived, least likely to have rehearsed a protocol, and most heavily documented in public audio and video. The blind spot does not belong to the twenty-year-old. It belongs to the household that has quietly delegated its security posture to them.

The Two Per Cent Problem

Return to the $563.1 million, and ask what kind of measurement it is.

In December 2025 the FTC published its annual report to Congress on protecting older consumers, and buried in the methodology is the most important number in American fraud policy. Drawing on work by the economist Devesh Raval, published in Marketing Science in 2020, which matched consumers named in nine FTC enforcement cases against complaints actually filed, the Commission adopted the assumption that Consumer Sentinel captures around 2.0 per cent of losses where the average loss is under $1,000, and around 6.7 per cent where it is over $1,000. On that basis, the FTC estimated that the true cost of fraud in the United States in 2024 could be as high as $195.9 billion, with between $10.1 billion and $81.5 billion of it borne by people aged 60 and over, depending on the method used.

Now apply the filter to the cohort in question. The median loss reported by twenty-somethings sits at $417, comfortably inside the band where roughly one loss in fifty is ever reported to anybody. The over-eighties, at a median of $1,650, sit in the band where roughly one in fifteen is. If those ratios are even approximately right, then the reporting apparatus systematically compresses the young end of the distribution far harder than the old end, and the gap between the $563.1 million and the $7.7 billion is, in reality, considerably narrower than the table implies.

To this add three further sources of invisibility peculiar to this age group. Reporting to the IC3 requires knowing you were defrauded, which victims of a well-executed impersonation frequently never learn. It requires believing the report will accomplish something, a belief that decays with each unanswered submission. And it requires overcoming embarrassment, which is heaviest precisely where the victim's self-image includes being too online to fall for it. The half-billion is not an exaggeration of a small problem. It is the visible fraction of a large one, seen through the coarsest filter in the dataset.

Whose Job Is Authentication

The advice at the end of every one of these warnings is sound and reads, in aggregate, like a sentence. Verify out of band. Ring back on a number you already had. Agree a family code word. Do not trust pictures or video sent to you. Make your profiles private. Assume urgency is an attack.

Consider what this costs. It is a proposal that every human being now perform manual cryptographic verification, unaided, on every unexpected contact, for the rest of their lives, using procedures they must design and maintain themselves. It converts every phone call from a family member into a challenge-response protocol. It is a tax on trust, levied per interaction, and it falls on individuals for a straightforward reason: the parties in a position to build authentication into the infrastructure have not been made to do it.

The regulatory record is thin. In February 2024 the Federal Communications Commission ruled that AI-generated voices in robocalls fall within the Telephone Consumer Protection Act's restrictions, a genuine and useful statement of principle that governs mass automated dialling rather than the targeted single call. The follow-on rulemaking, which proposed defining an AI-generated call and requiring disclosure at the start of one, had still not been finalised as a binding rule by September 2026. Britain went further on the money rather than the message, with mandatory reimbursement for authorised push payment fraud, and the results are instructive in both directions: UK Finance's 2026 annual report recorded £1.28 billion lost to payment fraud in 2025, with unauthorised fraud losses falling 5 per cent to £703.4 million while authorised push payment losses rose 19 per cent to £576.4 million. Making banks pay changes who absorbs the harm. It has not yet changed how much harm is manufactured.

Nor is the young cohort's exposure an American peculiarity. Research published by SquareTrade Europe on 2 September 2026, covering 15,000 adults across Europe including 3,000 in the United Kingdom, found that among its UK respondents 61 per cent of 25 to 34 year olds and 59 per cent of 18 to 24 year olds had been targeted by scams, against 20 per cent of the over-sixty-fives. Wherever you look, the people being approached most often are the people assumed to need the least protection.

The industry's own preferred answer arrived in June 2026, when the Aspen Institute's Financial Security Programme launched a Scam Prevention Initiative with a leadership group including JPMorganChase, Amazon, Apple, Google, Microsoft, PayPal, Walmart and Zelle. Kate Griffin, who directs it, said that fraud and scams are a shared threat and require a shared response. She is right. It is also worth noticing what a shared response is: an arrangement in which no single party is liable, measurement and convening are the deliverables, and the output is education. JPMorganChase runs more than a thousand fraud and scam workshops a year across its branches. Workshops are worth doing. They are also, compared with rebuilding call authentication or verifying consent before a voice can be cloned, extraordinarily cheap.

Which brings the argument back to the twenty-year-old, who is not the problem and never was. The digital native was never a category of person. It was a description of a birth year, dressed up as a competence, and the research community said so almost a decade ago and nobody in marketing, banking or education listened. What that cohort actually inherited was fluency with interfaces whose authentication was broken from the outset, plus a set of heuristics that detected the fingerprints of human labour in a fraud economy that no longer employs many humans. Both inheritances have now been withdrawn, one by generative text, the other by agents that can hold a conversation for the price of the electricity. The cruel part is not that they were overconfident. It is that the rest of the household believed them, wired its emergency verification through their judgement, and never mentioned that the machine on the other end had been built specifically to sound like someone they love.

Sources and References

  1. JPMorganChase, “Don't let scammers crash back-to-school season”, published via Mississippi Today, 21 August 2026. https://mississippitoday.org/2026/08/21/dont-let-scammers-crash-back-to-school-season/
  2. KPRC/Click2Houston, “Back-to-school scams targeting college students: How to protect yourself”, 1 September 2026. https://www.click2houston.com/news/local/2026/09/01/back-to-school-scams-targeting-college-students-how-to-protect-yourself/
  3. Internet Crime Complaint Center (FBI), “2025 IC3 Annual Report”, April 2026. https://www.ic3.gov/AnnualReport/Reports/2025_IC3Report.pdf
  4. Internet Crime Complaint Center (FBI), “2024 IC3 Annual Report”, April 2025. https://www.ic3.gov/AnnualReport/Reports/2024_IC3Report.pdf
  5. Federal Trade Commission, “Consumer Sentinel Network Data Book 2024”, March 2025. https://www.ftc.gov/system/files/ftc_gov/pdf/csn-annual-data-book-2024.pdf
  6. Federal Trade Commission, “Protecting Older Consumers 2024-2025: A Report of the Federal Trade Commission”, December 2025. https://www.ftc.gov/system/files/ftc_gov/pdf/P144400-OlderAdultsReportDec2025.pdf
  7. Devesh Raval, “Whose Voice Do We Hear in the Marketplace? Evidence from Consumer Complaining Behavior”, Marketing Science, vol. 39, no. 1, 2020, pp. 168-187. https://deveshraval.github.io/complaintBehavior.pdf
  8. Federal Trade Commission, “Reported losses to scams on social media eight times higher than in 2020”, Data Spotlight, 27 April 2026. https://www.ftc.gov/news-events/data-visualizations/data-spotlight/2026/04/reported-losses-scams-social-media-eight-times-higher-2020
  9. Federal Trade Commission, “Paying to get paid: gamified job scams drive record losses”, Data Spotlight, December 2024. https://search.ftc.gov/system/files/ftc_gov/pdf/task-scams-spotlight-2024.pdf
  10. Federal Trade Commission, “FTC Data Show People Reported Losing $3.5 Billion to Imposter Scams in 2025”, 15 June 2026. https://www.ftc.gov/news-events/news/press-releases/2026/06/ftc-data-show-people-reported-losing-3-point-5-billion-imposter-scams-2025
  11. Paul A. Kirschner and Pedro De Bruyckere, “The myths of the digital native and the multitasker”, Teaching and Teacher Education, vol. 67, October 2017, pp. 135-142. https://doi.org/10.1016/j.tate.2017.06.001
  12. Nature (editorial), “The digital native is a myth”, Nature, vol. 547, 27 July 2017, p. 380. https://doi.org/10.1038/547380a
  13. Dewan Mehrab Ashrafi, Mahnoor Sattar and Rafayat Jalal, “What influences digital natives' susceptibility to cyber phishing scams?”, Journal of Financial Crime, vol. 32, no. 6, December 2025, pp. 1211 onwards. https://doi.org/10.1108/JFC-01-2025-0021
  14. iProov, “Study Reveals Deepfake Blindspot: Only 0.1% of People Can Accurately Detect AI-Generated Deepfakes”, 11 February 2025. https://www.iproov.com/press/study-reveals-deepfake-blindspot-detect-ai-generated-content
  15. Fred Heiding, Simon Lermen, Andrew Kao, Bruce Schneier and Arun Vishwanath, “Evaluating Large Language Models' Capability to Launch Fully Automated Spear Phishing Campaigns: Validated on Human Subjects”, arXiv:2412.00586, December 2024. https://arxiv.org/abs/2412.00586
  16. Chengyi Wang, Sanyuan Chen, Yu Wu and colleagues (Microsoft), “Neural Codec Language Models are Zero-Shot Text to Speech Synthesizers”, arXiv:2301.02111, 5 January 2023. https://arxiv.org/abs/2301.02111
  17. McAfee, “Beware the Artificial Impostor”, McAfee Labs report, 2023. https://www.mcafee.com/content/dam/consumer/en-us/resources/cybersecurity/artificial-intelligence/rp-beware-the-artificial-impostor-report.pdf
  18. Sanket Badhe, “ScamAgents: How AI Agents Can Simulate Human-Level Scam Calls”, Proceedings of Machine Learning Research vol. 299 (Conference on Applied Machine Learning for Information Security), arXiv:2508.06457, 8 August 2025. https://arxiv.org/abs/2508.06457
  19. Xiangzhe Yuan, Jiajun Wang, Huanchen Wang, Qian Wan and Siying Hu, “ImmuniFraug: A Metacognitive Intervention Anti-Fraud Approach to Enhance Undergraduate Students' Cyber Fraud Awareness”, arXiv:2601.06774, 11 January 2026. https://arxiv.org/abs/2601.06774
  20. Karina LaRubbio, Alyssa Lanter, Seihyun Lee, Mahima Ramesh and Diana Freed, “Intergenerational Support for Deepfake Scams Targeting Older Adults”, arXiv:2508.11579, 15 August 2025. https://arxiv.org/abs/2508.11579
  21. Federal Communications Commission, “FCC Makes AI-Generated Voices in Robocalls Illegal”, 8 February 2024. https://www.fcc.gov/document/fcc-makes-ai-generated-voices-robocalls-illegal
  22. Kamlesh Harry, “UK Finance Fraud Report: adapting to the next phase of fraud and scams”, UK Finance, 17 July 2026. https://www.ukfinance.org.uk/news-and-insight/blog/uk-finance-fraud-report-adapting-next-phase-fraud-and-scams
  23. UK Finance, “Annual Fraud Report 2026”, June 2026. https://www.ukfinance.org.uk/system/files/2026-06/UK%20Finance%20Fraud%20Report%202026.pdf
  24. SecurityBrief UK, “Young Britons far more likely to be scammed online”, reporting SquareTrade Europe research, 2 September 2026. https://securitybrief.co.uk/story/young-britons-far-more-likely-to-be-scammed-online
  25. PYMNTS, “JPMorganChase and Amazon Back Aspen Institute Drive Against Scams”, 23 June 2026. https://www.pymnts.com/cybersecurity/fraud-prevention/2026/jpmorganchase-and-amazon-back-aspen-institute-drive-against-scams/

Tim Green

Tim Green UK-based Systems Theorist & Independent Technology Writer

Tim explores the intersections of artificial intelligence, decentralised cognition, and posthuman ethics. His work, published at smarterarticles.co.uk, challenges dominant narratives of technological progress while proposing interdisciplinary frameworks for collective intelligence and digital stewardship.

His writing has been featured on Ground News and shared by independent researchers across both academic and technological communities.

ORCID: 0009-0002-0156-9795 Email: tim@smarterarticles.co.uk

Listen to the free weekly SmarterArticles Podcast

Support What We Do

Discuss...