Valued at Millions, Compensated at Zero: The Human Data Exploitation Crisis

In February 2024, Reddit filed for an initial public offering and simultaneously announced a deal worth approximately $60 million per year granting Google access to its vast archive of user-generated conversations for the purpose of training artificial intelligence models. Reddit CEO Steve Huffman captured an emerging paradox of the digital age: “The source of artificial intelligence is actual intelligence. That's what you find on Reddit.” Within months, Reddit struck a similar arrangement with OpenAI, reportedly valued at around $70 million annually. In its IPO prospectus, Reddit disclosed that data licensing arrangements signed in January 2024 alone carried an aggregate contract value of $203 million over two to three years. The company's first earnings report as a public entity showed a 450 per cent year-over-year increase in non-advertising revenue, driven almost entirely by those licensing agreements.

Something strange had happened. A platform built on the unpaid contributions of millions of anonymous users had discovered that the messy, argumentative, sometimes brilliant, often profane corpus of human conversation it had accumulated over nearly two decades was now worth hundreds of millions of dollars. Not because advertisers wanted it. Because the machines needed it.

This is the paradox at the heart of artificial intelligence in 2026. AI systems can generate infinite synthetic content, flooding the internet with text, images, video and audio at a pace that dwarfs human output. Yet the data those systems need most, the human-created information that grounds their training and prevents their degradation, is becoming scarcer and more precious by the month. The implications for personal privacy and data rights are profound, unsettling, and largely unresolved.

The Photocopier Problem

In July 2024, a team of researchers led by Ilia Shumailov at the University of Oxford published a landmark paper in Nature demonstrating what happens when AI models train on the outputs of other AI models. The phenomenon, which the researchers termed “model collapse,” showed that large language models, variational autoencoders, and Gaussian mixture models all degrade when successive generations are trained on content produced by their predecessors. The tails of the original data distribution vanish first, eliminating rare events and minority perspectives. Eventually, the model's output bears little resemblance to the distribution of the real world it was supposed to represent.

Nicolas Papernot, an assistant professor of computer engineering at the University of Toronto and a co-author of the study, offered a vivid analogy. “A good analogy for this is when you take a photocopy of a piece of paper, and then you photocopy the photocopy,” he told the University of Toronto News. “Eventually, if you repeat that process many, many times, you will lose most of what was contained in that original piece of paper.” The research, published with collaborators from the Universities of Cambridge and Edinburgh and Imperial College London, found that training on AI-generated data not only degrades quality but further encodes the biases and errors already present in the training pipeline. Papernot warned that the findings “cast doubt on predictions that the current pace of development in LLM technology will continue unabated.” The paper received over 500 citations and an Altmetric attention score exceeding 3,600, reflecting the urgency with which the research community received its conclusions.

The timing of this discovery was particularly significant. By April 2025, a study by the SEO research firm Ahrefs, analysing 900,000 newly created web pages, found that 74 per cent contained AI-generated content. A separate analysis by the SEO firm Graphite, reported by Axios in May 2025, found that the share of newly published articles written by AI had reached approximately 52 per cent. Google search results containing AI-written pages climbed from 11 per cent in May 2024 to nearly 20 per cent by July 2025, according to an ongoing study by Originality.ai. An arXiv research paper from March 2025 estimated that at least 30 per cent of text on active web pages originates from AI-generated sources, with the actual proportion likely approaching 40 per cent. The internet is rapidly filling up with machine-generated text, and every drop of it threatens to contaminate the training pipelines of next-generation AI models.

This creates a vicious feedback loop. As AI-generated content proliferates, the proportion of authentic human-created data in any given web scrape declines. Models trained on this increasingly synthetic web produce outputs further removed from genuine human expression, which then get published and scraped again. Researchers at the International Conference on Learning Representations (ICLR) in 2025 found that this “strong model collapse” cannot generally be mitigated by simple data weighting adjustments. A separate paper at the International Conference on Machine Learning (ICML) in 2024 revealed that as synthetic data grows in training datasets, the traditional scaling laws that have driven AI progress begin to break down entirely.

The upshot is stark. The most valuable commodity in the AI economy is no longer processing power or algorithmic innovation. It is authentic, verified, human-generated data. And that realisation has set off a global scramble with enormous consequences for anyone who has ever posted, typed, spoken, or created anything online.

The Great Data Land Grab

The race to secure human data has produced a wave of licensing agreements that would have seemed improbable just a few years ago. The Associated Press was among the first major publishers to sign a deal with OpenAI in July 2023, granting access to its news archive dating back to 1985. Google struck its first AI content licensing agreement with the AP in January 2025. The Financial Times signed a content licensing deal with OpenAI in April 2024. News Corp agreed to a multi-year arrangement reportedly worth up to $250 million over five years. Conde Nast and Time also entered agreements. By early 2025, the wave had reached The Guardian, The Washington Post, Axios, and the Norwegian publisher Schibsted Media.

These deals represent a fundamental shift in the economics of content creation. For decades, digital publishers watched their revenues erode as platforms aggregated their content and captured the advertising value. Now, the same dynamic is playing out again, but with a new twist: the platforms are not just displaying human content to attract eyeballs. They are consuming it to build intelligence. And this time, at least some publishers are negotiating payment.

But the deals also expose a deeper asymmetry. The individuals who actually created the content receive nothing directly. The Reddit users whose posts are now worth $60 million a year to Google, the journalists whose reporting trains ChatGPT, the photographers whose images teach image generators to see, are not party to any of these agreements. Huffman himself acknowledged this tension in a 2024 interview with Fast Company: “As more content on the internet is written by machines, there's an increasing premium on content that comes from real people.” Reddit, he noted, has “nearly two decades of authentic conversation” and more than 16 billion comments. The premium is real. The compensation flows to the platform, not to the people who made the platform valuable.

Reddit has also aggressively defended its data from unauthorised extraction. After years of being “scraped every which way,” as Huffman put it, the company updated its robots.txt file in July 2024 to block all web crawlers except Google. Huffman publicly accused Microsoft of training its AI services on Reddit data “without telling us,” and named Anthropic and Perplexity as companies that had also trained their systems using Reddit content without permission. In late 2025, Reddit filed lawsuits against Perplexity AI and Anthropic. The company has since proposed a “dynamic pricing” model for its data, seeking compensation that increases as its content becomes more essential to AI-generated answers, rather than accepting fixed licensing fees.

This dynamic echoes a framework articulated by Shoshana Zuboff, the Harvard Business School professor emerita whose 2019 book The Age of Surveillance Capitalism described the extraction of human behavioural data as the defining feature of the digital economy. Zuboff argued that technology companies had claimed “human experience as free raw material for hidden commercial practices of extraction, prediction, and sales.” The AI training data economy takes this logic and intensifies it. Where surveillance capitalism extracted behavioural surplus from user interactions to predict future actions, the new data economy extracts the creative and intellectual output itself, using it not merely to predict behaviour but to replicate and replace the capabilities of its creators.

When Deletion Becomes Impossible

The rising value of human data collides directly with one of the foundational principles of modern privacy law: the right to be forgotten. Under Article 17 of the European Union's General Data Protection Regulation, individuals have the right to request the erasure of their personal data. California's landmark AB 1008, signed into law by Governor Gavin Newsom in September 2024 and effective from January 2025, went further still, amending the California Consumer Privacy Act to specify that personal information can exist in “abstract digital formats,” including “artificial intelligence systems that are capable of outputting personal information.” Under this law, consumers have the right to access, delete, correct, and restrict the sale of personal data contained within trained AI systems, including data encoded in tokens or model weights. California also passed SB 1223 alongside AB 1008, introducing neural data as a category of sensitive personal information subject to even stricter protections.

The problem is that complying with these rights is, at present, somewhere between extraordinarily difficult and functionally impossible. AI models do not store information in discrete, retrievable entries the way a database does. Once personal data has been absorbed into a model's parameters through the training process, it is distributed across billions of numerical weights in ways that cannot be straightforwardly traced or extracted. Personal data can appear in multiple layers of the AI stack: raw training datasets, tokenised text, embeddings, model checkpoints, and fine-tuned weights. As one expert quoted by MIT Technology Review observed, “You can assume that any large-scale web-scraped data always contains content that shouldn't be there.”

The European Data Protection Board acknowledged this challenge in a January 2025 technical report, stating that the right to erasure requires reversing the “memorisation of personal data by the model,” involving deletion of both “the personal data used as input for training” and “the influence of that data on the model.” The Board has made the right to erasure an enforcement priority for 2025, with 32 Data Protection Authorities across Europe participating in coordinated investigations.

The emerging field of “machine unlearning” attempts to address this gap, but the technology remains immature. Exact unlearning methods, such as the SISA framework, require partitioning training datasets and retraining from earlier checkpoints. Approximate methods aim to selectively remove the influence of specific data points without full retraining. But there is no universally accepted standard for verifying whether unlearning has been effective. As a November 2025 research paper from the Centre for Emerging Policy noted, machine unlearning methods “have been there for several years but have not been put into industry practice, which reflects the immaturity of this stream of methods.” Engineers acknowledge that the only truly reliable method of removing an individual's data from a model is to retrain it from scratch, a process costing millions of dollars and weeks of computation time for frontier models.

The practical reality in 2026 is that the right to erasure operates primarily at the input and output layers, not within the model itself. Companies can delete source training data and implement output filters to prevent models from generating specific personal information. But the influence of that data on the model's learned parameters persists. The Hamburg Data Protection Authority has argued that large language models do not store personal data in a way that triggers data protection obligations. Other authorities disagree sharply. The GDPR itself contains exceptions that further complicate compliance, allowing companies to deny erasure requests on grounds including archiving in the public interest and scientific research, providing potential justification for retaining training data even when individuals demand its removal.

For individuals, the implications are deeply concerning. The more valuable human data becomes, the greater the incentive for companies to acquire, retain, and resist deleting it. And the technical architecture of modern AI makes meaningful erasure a problem that legal frameworks have not yet solved.

Synthetic Abundance and Its Discontents

The mirror image of human data scarcity is synthetic data abundance. The synthetic data generation market, valued at approximately $400 million to $500 million in 2025 according to Mordor Intelligence and Grand View Research, is projected to reach between $2 billion and $9 billion by the end of the decade, with growth rates ranging from 25 to 46 per cent annually. In March 2025, NVIDIA acquired the synthetic data startup Gretel for more than $320 million, integrating its privacy-preserving data generation platform into its AI development tools. Gretel's technology allows organisations to generate realistic datasets that retain the statistical properties of real-world data while ensuring no actual personal information is disclosed.

The appeal of synthetic data for privacy is obvious. If AI models can be trained on data that was never derived from real individuals, many of the thorniest privacy and consent problems simply evaporate. The EU AI Act, fully applicable from 2 August 2026, explicitly establishes a hierarchy in which synthetic and anonymised data should be used before processing sensitive personal data. Article 10(5) specifies that providers of high-risk AI systems may only process special categories of personal data for bias detection and correction if the goal “cannot be effectively fulfilled by processing synthetic or anonymised data.”

Yet synthetic data brings its own considerable risks. The model collapse research demonstrates that over-reliance on synthetic training data degrades model quality over successive generations. Gartner has predicted that by 2027, 60 per cent of data and analytics leaders will face critical failures in managing synthetic data, risking AI governance, model accuracy, and compliance. Synthetic data may be privacy-preserving in principle, but it is not a substitute for the diversity, unpredictability, and grounding in lived experience that human-generated data provides.

The Epoch AI research group has documented the scale of the problem. The total effective stock of human-generated public text data amounts to roughly 300 trillion tokens, with an 80 per cent confidence interval suggesting this stock will be fully utilised for AI training sometime between 2028 and 2032. Pablo Villalobos, lead author of Epoch's study “Will we run out of data? Limits of LLM scaling based on human-generated data,” has acknowledged that “some relatively small but very high-quality sources have not been tapped yet,” including digitised documents in libraries, but warned that dwindling reserves “might not be enough” to postpone the issue significantly. OpenAI researchers have confirmed that during the development of GPT-4.5, a shortage of fresh data was more of a constraint than a lack of computing power.

The scarcity of human data and the abundance of synthetic data create a peculiar economic inversion. In the data broker market, valued at $303 billion to $333 billion in 2025, the average cost of personal data for an individual aged 18 to 25 is just $0.36, according to VPNCentral. For those over 55, it falls to $0.05. These figures reflect the commoditised value of personal data in the advertising economy. But in the AI training economy, the same human data takes on an entirely different character. It is not purchased per record from a broker. It is licensed in bulk, for millions of dollars, from platforms that aggregated it. The value of your data is simultaneously trivial and enormous. You are paid for neither.

The Question of Data Dignity

This asymmetry has revived interest in a concept first articulated by Jaron Lanier and E. Glen Weyl in their 2018 Harvard Business Review essay “A Blueprint for a Better Digital Society.” Lanier and Weyl proposed the idea of “data dignity,” arguing that data generated through interactions with digital systems constitutes a form of labour that should be compensated. They envisioned organisations called “mediators of individual data,” or MIDs, functioning as unions for data contributors. These MIDs would negotiate collectively with technology companies over access, usage, and royalties.

The concept remained largely theoretical until generative AI made the exploitation of human creative output visible at an industrial scale. Artists, writers, musicians, and photographers discovered that their work had been scraped from the internet and fed into training datasets without consent or compensation. Reddit users learned their posts were training chatbots. Authors found their books in the Books3 dataset. Photographers recognised their images in the outputs of image generators. The discovery was not that data had value. It was that the people who created it had been systematically excluded from capturing any of that value.

The “data as labour” framework has gained renewed academic attention. A paper published in Business Ethics Quarterly examined the labour analogy in depth, arguing that if data contributions are “characterised by asymmetric bargaining power of the kind found in the labour market, we should embrace proposals such as the creation of data unions and data strikes and similar collective actions by data contributors.” The American Economic Association has published research on the concept, arguing that treating data as capital “neglects users' roles in creating data, reducing incentives for users, distributing the gains from the data economy unequally, and stoking fears of automation.”

Yet the data dignity framework has its critics. The communications theorist Nick Couldry has suggested that paying people for their data may actually undermine rather than enhance human dignity, by “further commodifying our lives, treating us as mere labourers or passive resources to be mined.” If the solution to the exploitation of human data is to make that exploitation transactional, have we resolved the problem or merely normalised it?

The Regulatory Scramble

Legislators and regulators around the world are grappling with these questions, but responses remain fragmented and often contradictory.

The European Union's AI Act represents the most comprehensive legislative attempt to govern AI and data. Fully applicable from August 2026, it imposes strict requirements on data governance for high-risk AI systems, mandating that training data be relevant, representative, and accompanied by documentation of collection methods. Non-compliance carries penalties of up to 35 million euros or 7 per cent of global annual turnover. Transparency obligations for general-purpose AI model providers, including requirements to disclose copyrighted training data, took effect in August 2025.

In the United States, the landscape is more fractured. California's AB 1008 is the most ambitious state-level effort, explicitly extending privacy rights into AI model weights. Colorado's Algorithmic Accountability Law, effective February 2026, grants consumers rights to notice, explanation, correction, and appeal for high-risk AI decisions. But there is no federal data protection law. David Evan Harris, who teaches AI ethics at UC Berkeley, has described this gap as leaving Americans with “no standardised legal right to opt out of AI training.” Marietje Schaake, international policy director at Stanford's Cyber Policy Centre, has observed: “We have the GDPR in Europe, we have the CCPA in California, but there's still no federal data protection law in America.”

The 47th Global Privacy Assembly, held in Seoul in September 2025 and attended by over 140 authorities from more than 90 countries, adopted a resolution noting that “the public availability of personal data does not automatically imply a lawful basis for its processing” for AI training purposes. France's CNIL has been particularly active, publishing recommendations urging AI developers to incorporate privacy protection from the design stage.

In the United Kingdom, the approach has been characteristically principles-based. The Financial Conduct Authority confirmed in September 2025 that it would not introduce AI-specific regulations. The Competition and Markets Authority, armed with new powers under the Digital Markets, Competition and Consumers Act 2024, can now investigate breaches of consumer protection law directly and impose fines of up to 10 per cent of global turnover. Whether these powers will be used to address the extraction of personal data for AI training remains to be seen.

The Emerging Stratification of Data

The convergence of model collapse, data scarcity, privacy regulation, and the rising economic value of authentic human content is producing a new stratification of information. At the top sit curated, high-quality datasets licensed from publishers, platforms, and institutions. These command premium prices and form the foundation of frontier AI models. In the middle sits synthetic data, cheap and abundant but requiring careful curation to avoid degrading model performance. At the bottom sits the vast, unsorted mass of web-scraped content, increasingly contaminated by AI-generated material and of diminishing value for training purposes.

This hierarchy has implications for power as well as privacy. The organisations that control large repositories of authentic human data occupy a position of increasing strategic importance. Reddit understood this early, monetising its user base not through advertising alone but through the licensing of its conversational corpus. The question is whether the individuals whose contributions created that corpus will ever share in the value it generates.

Tamay Besiroglu, a co-author of the Epoch AI study on data depletion, compared the situation to “a literal gold rush” that depletes finite natural resources, warning that the AI field might face challenges in maintaining its current pace of progress once it drains the reserves of human-generated writing. If that projection proves correct, the organisations that have already secured exclusive access to high-quality human data will possess an advantage that is difficult to replicate.

For ordinary individuals, this future raises uncomfortable questions. Every social media post, every product review, every comment thread contributes to a collective resource that is being enclosed and monetised by corporations. The privacy frameworks designed to protect personal data were built for an era of databases and profiles, not for an era in which the very patterns of human thought and expression have become the raw material of a trillion-dollar industry.

What Authentic Expression Is Actually Worth

Stanford's Institute for Human-Centred Artificial Intelligence has proposed a shift from opt-out to opt-in data sharing, arguing that the default should be that data is not collected unless individuals affirmatively allow it. The precedent is instructive: when Apple introduced App Tracking Transparency in 2021, requiring apps to request permission before tracking users, industry estimates suggest that 80 to 90 per cent of people chose not to allow tracking. If a similar opt-in framework were applied to AI training data, the supply of available human data would contract dramatically, further increasing its scarcity value and the incentive to either circumvent consent mechanisms or develop viable synthetic alternatives.

Cisco's 2025 Data Privacy Benchmark Study found that 64 per cent of respondents worry about inadvertently sharing sensitive information with generative AI tools. That concern is not unfounded. A California lawsuit filed in 2025 accuses Google's Gemini of accessing users' private communications, alleging that a policy change gave the chatbot default access to private content such as emails and attachments, reversing a previous opt-in model. Technology companies, as Al Jazeera reported in November 2025, are “rarely fully transparent about the user data they collect and what they use it for.”

The tension between privacy and utility is not new, but AI has sharpened it beyond recognition. Privacy advocates argue that individuals should have meaningful control over how their data is used, including the right to withdraw it from AI training pipelines. AI developers counter that the technology cannot advance without access to diverse, representative human data, and that restricting access will entrench the dominance of companies that have already amassed large datasets. Both arguments contain truth, and neither resolves the fundamental question: in an economy where human creativity and expression have become the most valuable raw material for machine intelligence, who should decide how that material is used, and who should benefit from its exploitation?

The answer will not emerge from a single regulation, technology, or market mechanism. It will require a renegotiation of the relationship between individuals, platforms, and the AI systems that increasingly mediate our experience of the world. The data we generate is not merely a commodity to be bought and sold. It is an expression of who we are, how we think, and what we value. In the age of synthetic abundance, human data is not becoming less important. It is becoming more important, more contested, and more urgently in need of protection. The machines can generate infinite content. But they cannot generate meaning. That still comes from us. And until we collectively decide what that is worth, the value will continue to accrue to those who have the infrastructure to extract it.

References & Sources

  1. Shumailov, I. et al. (2024). “AI models collapse when trained on recursively generated data.” Nature, 631, 755-759. https://doi.org/10.1038/s41586-024-07566-y

  2. University of Toronto. (2024). “Training AI on machine-generated text could lead to 'model collapse.'” https://www.utoronto.ca/news/training-ai-machine-generated-text-could-lead-model-collapse-researchers-warn

  3. PBS NewsHour. (2024). “AI 'gold rush' for chatbot training data could run out of human-written text as early as 2026.” https://www.pbs.org/newshour/economy/ai-gold-rush-for-chatbot-training-data-could-run-out-of-human-written-text-as-early-as-2026

  4. Villalobos, P. et al. (2022). “Will we run out of data? Limits of LLM scaling based on human-generated data.” arXiv:2211.04325.

  5. Ahrefs. (2025). “74% of New Webpages Include AI Content.” https://ahrefs.com/blog/what-percentage-of-new-content-is-ai-generated/

  6. Axios. (2025). “AI-written web pages haven't overwhelmed human-authored content, study finds.” https://www.axios.com/2025/10/14/ai-generated-writing-humans

  7. Originality.ai. (2025). “Amount of AI Content in Google Search Results.” https://originality.ai/ai-content-in-google-search-results

  8. Fast Company. (2024). “CEO Steve Huffman on Reddit's essential humanity in the AI era.” https://www.fastcompany.com/90997770/reddit-steve-huffman-interview-ai-ipo-2024

  9. TechCrunch. (2024). “Reddit says it's made $203M so far licensing its data.” https://techcrunch.com/2024/02/22/reddit-says-its-made-203m-so-far-licensing-its-data/

  10. Columbia Journalism Review. (2025). “Reddit Is Winning the AI Game.” https://www.cjr.org/analysis/reddit-winning-ai-licensing-deals-openai-google-gemini-answers-rsl.php

  11. Digiday. (2024/2025). Timelines of publisher-AI licensing deals. https://digiday.com/media/

  12. Zuboff, S. (2019). The Age of Surveillance Capitalism. PublicAffairs.

  13. Securiti. (2024). “AB 1008: California's Move to Regulate AI and Personal Data.” https://securiti.ai/blog/california-ab-1008-to-regulate-ai-and-personal-data/

  14. Future of Privacy Forum. (2025). “Do LLMs Contain Personal Information?” https://fpf.org/blog/do-llms-contain-personal-information-california-ab-1008-highlights-evolving-complex-techno-legal-debate/

  15. European Data Protection Board. (2025). “Effective implementation of data subjects' rights.” https://www.edpb.europa.eu/system/files/2025-01/d2-ai-effective-implementation-of-data-subjects-rights_en.pdf

  16. TechPolicy.Press. (2025). “The Right to Be Forgotten Is Dead: Data Lives Forever in AI.” https://www.techpolicy.press/the-right-to-be-forgotten-is-dead-data-lives-forever-in-ai/

  17. CEP Project. (2025). “Machine Unlearning and the Right to be Forgotten Under Emerging Legal Frameworks.” https://cep-project.org/

  18. Mordor Intelligence. (2025). “Synthetic Data Market Report.” https://www.mordorintelligence.com/industry-reports/synthetic-data-market

  19. TechCrunch. (2025). “Nvidia reportedly acquires synthetic data startup Gretel.” https://techcrunch.com/2025/03/19/nvidia-reportedly-acquires-synthetic-data-startup-gretel/

  20. EU AI Act. (2024). Regulation (EU) 2024/1689. https://artificialintelligenceact.eu/

  21. Gartner. (2025). “Top Data & Analytics Predictions.” https://www.gartner.com/en/newsroom/press-releases/2025-06-17-gartner-announces-top-data-and-analytics-predictions

  22. Lanier, J. and Weyl, E.G. (2018). “A Blueprint for a Better Digital Society.” Harvard Business Review.

  23. Cambridge Core. (2024). “Is Data Labor?” Business Ethics Quarterly. https://www.cambridge.org/core/journals/business-ethics-quarterly/

  24. Stanford HAI. (2025). “Privacy in an AI Era.” https://hai.stanford.edu/news/privacy-ai-era-how-do-we-protect-our-personal-information

  25. Future of Privacy Forum. (2025). “GPA 2025: AI development and human oversight.” https://fpf.org/blog/gpa-2025-ai-development-and-human-oversight-of-decisions-involving-ai-systems-were-this-years-focus-for-global-privacy-regulators/

  26. VPNCentral. (2025). “Data Brokering Statistics.” https://vpncentral.com/data-brokering-statistics/

  27. Cisco. (2025). “2025 Data Privacy Benchmark Study.”

  28. ICLR. (2025). “Strong Model Collapse.” https://proceedings.iclr.cc/

  29. MIT Technology Review. (2025). “A major AI training data set contains millions of examples of personal data.” https://www.technologyreview.com/2025/07/18/1120466/

  30. Al Jazeera. (2025). “Are tech companies using your private data to train AI models?” https://www.aljazeera.com/news/2025/11/24/

  31. American Economic Association. (2018). “Should We Treat Data as Labor?” https://www.aeaweb.org/articles?id=10.1257/pandp.20181003

  32. CNIL. (2025). “AI and GDPR: new recommendations.” https://www.cnil.fr/en/ai-and-gdpr-cnil-publishes-new-recommendations-support-responsible-innovation

  33. WebProNews. (2025). “Reddit's Billion-Dollar Bet.” https://www.webpronews.com/

  34. World Economic Forum. (2025). “Artificial intelligence and the growth of synthetic data.” https://www.weforum.org/stories/2025/10/ai-synthetic-data-strong-governance/

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

Discuss...