Late Lessons, Jensen Huang and AI
Markdown version

Andrew D. Maynard: scholarly and long-form work on risk, emerging technologies, responsible innovation and AI#

Compiled 26 September 2026 for the Late Lessons × AI project.

What we already hold: the Substack corpus (working/maynard/corpus/, 391 posts, 2014–2026; only about 80 posts date from before 2023) and Films from the Future (2018) in Resources/. The full EEA Late lessons from early warnings II (2013) is also in Resources/, and it contains Maynard’s co-authored nanotechnology chapter.

Sources searched: - OpenAlex author A5090218914: 288 works, with many misattributions (see section c). - ORCID 0000-0003-2117-5128: 171 works. - Crossref author queries, filtered by name. - arXiv API and Zenodo API. - Google Scholar profile b8NhWc4AAAAJ. - andrewmaynard.net: the preprints page, the WordPress posts, pages and media APIs, and the academic CV dated 4 August 2026. - The 2020science.org archive API. - Nature article standfirsts, for the paywalled columns. - Web search.

How coverage was checked: each item’s title and key terms were grepped against the corpus. For cross-posted essays I also checked verbatim snippets.

Coverage labels: - FULL: the corpus holds the text or a near-complete adaptation. - PARTIAL: the corpus summarises or discusses the substance. - MENTION: the corpus only cites or links the item. - NONE: the item is not in the corpus. - IN RESOURCES: the item is already in Resources/.

Exclusions: - Modem Futura podcast episodes, as the user asked. Other podcasts (Mission: Interplanetary, Future Out Loud) are also left out. - Risk Bites videos and media interviews. - Maynard’s pure aerosol-physics and toxicology papers, unless they are landmark risk items.

(a) Priority list: what Andrew should add to Resources/#

HIGH priority#

These are central to his thinking on risk and AI, and the corpus does not hold their substance.

  1. The April 2026 synthesis essays on andrewmaynard.net. None of them is in the Substack corpus. They are free web pages; save each as text or PDF. - “What Thirty Years of Emerging Technology Risks Taught Me About Artificial Intelligence” (12 Apr 2026, about 2,200 words). https://andrewmaynard.net/2026/04/12/ai-and-the-future-of-being-human/ - “What Nanotechnology Taught Me About Governing AI” (12 Apr 2026, about 1,700 words). https://andrewmaynard.net/2026/04/12/ai-governance-risk-innovation/ - “Honest Non-Signals, Constitutive Resonance, and the Frameworks We Need for Understanding Human-AI Interaction” (12 Apr 2026, about 2,400 words). https://andrewmaynard.net/2026/04/12/ai-theory-cognitive-trojan-horse-constitutive-resonance/ - “The Future We’re Building, Whether We Mean To or Not” (12 Apr 2026, about 2,200 words). https://andrewmaynard.net/2026/04/12/ai-humanity-and-the-future/ - “The Three S-Curves: What AI Is Actually Doing in Higher Education” (12 Apr 2026, about 1,800 words). https://andrewmaynard.net/2026/04/12/ai-and-education-three-s-curves/ - “Before the Fourth Industrial Revolution: Notes on an Institutional Prehistory” (8 Apr 2026, about 2,500 words). https://andrewmaynard.net/2026/04/08/fourth-industrial-revolution-prehistory-wef-councils/ - Lower value within the same set: “What Happens to Books When AI Becomes the Reader?” and “Stick Figures, Sci-Fi Movies, and the Obligation to Make AI Accessible” (both 12 Apr 2026).

Why they matter: these are his own retrospective synthesis of 30 years of emerging-technology risk work applied to AI. They cover his congressional testimony, the NNI and the Project on Emerging Nanotechnologies (PEN), lessons from GMOs, why AI governance is repeating the “leave it to the experts” mistake, and the WEF early-warning institution proposals. That is the closest thing in his output to “Late Lessons for AI”, and none of it is in the corpus.

  1. Hansen, S. F., Maynard, A. D., Baun, A. & Tickner, J. A. (2008). “Late lessons from early warnings for nanotechnology.” Nature Nanotechnology 3(8): 444–447. doi:10.1038/nnano.2008.198. Closed access; Andrew should have the PDF. - What it is: a systematic check of nanotechnology against each of the EEA 2001 report’s twelve late lessons. It is his first direct engagement with the EEA framework, so it links his risk thinking to both EEA reports. - Corpus coverage: MENTION only. The 2018-12-15 Luddite post cites the EEA reports but not this paper. - Add alongside it the free 716-word blog post that introduced the paper: “Late lessons from early warnings”, 2020science.org, 20 July 2008, https://2020science.org/2008/07/20/8720-late-lessons-from-early-warnings/ (not in the corpus). - The 2013 EEA follow-on chapter by the same team is already in Resources/; see section (b).

  2. Maynard, A. D. & Dudley, S. M. (2023). “Navigating advanced technology transitions: using lessons from nanotechnology.” Nature Nanotechnology 18(10): 1118–1120. doi:10.1038/s41565-023-01481-5. Closed access. - What it is: his explicit, peer-reviewed bridge from nanotechnology governance lessons to AI, DNA technologies and quantum technologies. - Corpus coverage: PARTIAL. The only coverage is a 670-word pointer post, responsible-ai-lessons-from-nanotechnology (2023-10-02). - Add alongside it the free companion article: Maynard & Dudley (2023), “Navigating the risks and benefits of AI: Lessons from nanotechnology on ensuring emerging technologies are safe as well as successful”, The Conversation, https://theconversation.com/navigating-the-risks-and-benefits-of-ai-lessons-from-nanotechnology-on-ensuring-emerging-technologies-are-safe-as-well-as-successful-210872

  3. The “Safe handling” pair: Andrew’s own early-warning episode and his retrospective on it. Corpus coverage: NONE for both. - Maynard, A. D. & Aitken, R. J. (2016). “‘Safe handling of nanotechnology’ ten years on.” Nature Nanotechnology 11(12): 998–1000. doi:10.1038/nnano.2016.270. Closed access. - Maynard, A. D., Aitken, R. J., Butz, T., Colvin, V., Donaldson, K., Oberdörster, G., Philbert, M. A., Ryan, J., Seaton, A., Stone, V., Tinkle, S. S., Tran, L., Walker, N. J. & Warheit, D. B. (2006). “Safe handling of nanotechnology.” Nature 444: 267–269. doi:10.1038/444267a. Free to read at nature.com; about 1,500 citations.

Why they matter: the 2006 paper set five “grand challenges” for nano safety, and the 2016 paper audits what happened to them ten years later. Together they are a first-hand case study of how the early warnings on a technology were acted on, or not, and they map directly onto the EEA framing.

  1. Maynard, A. D. (2026). “The AI Cognitive Trojan Horse: How Large Language Models May Bypass Human Epistemic Vigilance.” arXiv:2601.07085, v2 revised 26 May 2026. doi:10.48550/arXiv.2601.07085. Open access. - What it is: his main statement on the cognitive and epistemic risk of AI. It covers “honest non-signals”, four bypass mechanisms and testable predictions. - Corpus coverage: PARTIAL. The corpus has the pre-paper essay is-ai-a-cognitive-trojan-horse (2026-01-10) and a process post, i-cracked-and-wrote-an-academic-paper (2026-01-17). The paper itself and the v2 revisions are not in the corpus.

  2. Maynard, A. D. (2026). “Constitutive Resonance as a Novel Framework for Understanding and Navigating Human-AI Interactions.” SSRN 6343880. doi:10.2139/ssrn.6343880. The preprint v3 PDF is free at andrewmaynard.net/wp-content/uploads/2026/03/constitutive_resonance_maynard_preprint_v3.pdf. - What it is: his theory of how conversational AI enters human self-constitution. It is core to his “future of being human” and cognition strand. - Corpus coverage: MENTION only, in magnifica-humanitas-and-being-human (2026-05-21).

  3. Maynard, A. D. (2026). “Constituting Responsibility: What Constitutional AI Reveals About the Limits and Futures of Responsible Innovation.” Zenodo. doi:10.5281/zenodo.22256794. v1.1 PDF: andrewmaynard.net/wp-content/uploads/2026/09/Constituting_Responsibility_V1.1.pdf. Open access. - What it is: his most direct scholarly treatment of whether the responsible-innovation framework works for frontier AI. It covers the “internalization problem” and inclusion deficits. - Corpus coverage: NONE. No dedicated post exists. - Note: Google Scholar lists Claude as a co-author of this paper.

  4. Abbott, J. & Maynard, A. (2025). AI and the Art of Being Human: A practical guide to thriving with AI while rediscovering yourself in the process. Waymark Works Publishing. ISBN 9798993145303. The Pocket Edition followed in 2026. - Access: free AI-readable full versions at https://www.aiandtheartofbeinghuman.com/ai-companion. There is also a free 50-page preview PDF on futureofbeinghuman.asu.edu. - Corpus coverage: PARTIAL. Only launch and meta posts are in the corpus: heads-up-on-new-ai-book, ai-and-the-art-of-being-human, ai-resurrecting-deceased-darlings (the missing foreword), 21-tools-for-thriving-with-ai, why-were-giving-away-our-book-on-thriving-with-ai and the-ai-book-i-actually-carry-with. The book text is not.

MEDIUM priority#

These are important background. The corpus holds only a popular or partial version.

  1. The core statements of risk innovation. He defines risk as a “threat to value” and introduces “orphan risks”. - Maynard (2015), “Why we need risk innovation”, Nature Nanotechnology 10: 730–731, doi:10.1038/nnano.2015.196. Coverage: PARTIAL, through the popular version thinking-innovatively-about-the-risks-of-tech-innovation (2016-01-11), plus risk-innovation-and-the-future (2020-11-05). - Maynard (2016), “Navigating the risk landscape”, Nature Nanotechnology 11: 211–212, doi:10.1038/nnano.2016.28. Coverage: NONE. - Maynard (2017), “Rethinking Risk”, in E. Finn & J. Eschrich (eds), Visions, Ventures, Escape Velocities: A Collection of Space Futures, ASU Center for Science and the Imagination. Free at csi.asu.edu/books/vvev. Coverage: PARTIAL, as an excerpt in tech-companies-need-a-social-risk-reboot (2018-09-03). - The two Nature Nanotechnology columns are closed access.

  2. Maynard (2015). “Navigating the fourth industrial revolution.” Nature Nanotechnology 10: 1005–1006. doi:10.1038/nnano.2015.286. Closed access.

    • What it is: his most-cited governance piece (268 citations), on converging technologies.
    • Corpus coverage: PARTIAL, only through the 2016 post criticising what Klaus Schwab leaves out. The April 2026 “prehistory” essay in item 1 gives the backstory.

  3. The WEF early-warning institution proposals. Maynard drafted these with Tim Harper for the Global Agenda Council on Emerging Technologies. They are a real proposal for an emerging-technology early-warning body.

    • The 2008 one-page idea, “Global Institute on Emerging Technology Policy”: andrewmaynard.net/wp-content/uploads/2023/12/Emerging-Tech-breakthrough-idea_Maynard.pdf
    • The 2010 proposal for a “Global Centre for Emerging Technology Intelligence” (CETI), published in WEF’s Everybody’s Business report: andrewmaynard.net/wp-content/uploads/2023/12/CETI_GRI_Published.pdf
    • Maynard & Harper (2011), Building a Sustainable Future, a WEF white paper.
    • Corpus coverage: MENTION only, in un-governing-ai-for-humanity (2023-12-22).

  4. Nanotechnology regulation lessons that parallel AI debates. Corpus coverage: NONE for all four.

    • Hodge, G. A., Maynard, A. D. & Bowman, D. M. (2014), “Nanotechnology: Rhetoric, risk and regulation”, Science and Public Policy 41(1): 1–14, doi:10.1093/scipol/sct029.
    • Maynard (2011), “Don’t define nanomaterials”, Nature 475: 31, doi:10.1038/475031a. On definitions as regulatory triggers, which parallels AI Act-style definitions.
    • Maynard, Bowman & Hodge (2011), “The problem of regulating sophisticated materials”, Nature Materials 10: 554–557, doi:10.1038/nmat3085.
    • Maynard & Rejeski (2009), “Too small to overlook”, Nature 460: 174, doi:10.1038/460174a. On the failure of voluntary industry reporting, which parallels voluntary AI commitments.
    • All four are closed access.

  5. Hansen, S. F., Maynard, A., Baun, A., Tickner, J. A. & Bowman, D. M. (2014). “What are the warning signs that we should be looking for?” In M. Hull & D. Bowman (eds), Nanotechnology Environmental Health and Safety, 2nd ed., Elsevier, pp. 9–24. doi:10.1016/B978-1-4557-3188-6.00002-5. Closed access.

    • What it is: the early-warnings team’s follow-on to the EEA chapter.
    • Corpus coverage: NONE.

  6. Maynard, A. (2020). Future Rising: A Journey from the Past to the Edge of Tomorrow. Mango Publishing. Andrew holds the text. Free excerpts: andrewmaynard.net/wp-content/uploads/2024/05/Future-Rising-Part-I.pdf and …/2024/12/Future-Rising-Intrduction.pdf.

    • What it is: his big-picture book on the future and complexity.
    • Corpus coverage: MENTION or meta only: eight-things-about-future-rising… (2020), future-rising-short-history-of-tomorrow (2024) and codesignal-future-rising (2025).

  7. Maynard (2026). “What the Rapid Adoption of the ‘Harness’ Metaphor in Artificial Intelligence Reveals About How We Conceptualize Human–AI Relations.” SSRN 6352678. doi:10.2139/ssrn.6352678. Open access.

    • Corpus coverage: PARTIAL. The essay version is what-we-miss-when-we-talk-about-ai-harnesses (2026-02-22). The paper adds the metaphor theory and the automation-bias argument.

  8. Maynard, A. & Stilgoe, J. (2020). “Introduction: The ethics of noumenal technologies.” In Maynard & Stilgoe (eds), The Ethics of Nanotechnology, Geoengineering and Clean Energy, Routledge Library of Essays on the Ethics of Emerging Technologies, pp. 1–13. doi:10.4324/9781003075028-1. Closed access.

    • What it is: his concept of technologies whose workings and effects are hidden. It bears on AI opacity.
    • Corpus coverage: NONE.

  9. Maynard, A. D., Grobe, A. & Renn, O. (2013). “Responsible innovation, global governance, and emerging technologies.” In R. Parker & R. Appelbaum (eds), Can Emerging Technologies Make a Difference in Development?, Routledge, pp. 168–187. doi:10.4324/9780203135686-21. Closed access.

    • Corpus coverage: NONE.

  10. Wang, J., Maynard, A. D., Lobo, J., Michael, K., Motsch, S. & Strumsky, D. (2024). “Knowledge Combination Analysis Reveals That Artificial Intelligence Research Is More Like ‘Normal Science’ Than ‘Revolutionary Science’.” Proceedings of HICSS-57. doi:10.24251/hicss.2024.673. Open access (hdl:10125/107058).

    • Why it matters: it gives empirical grounds for questioning claims that AI is revolutionary.
    • Corpus coverage: NONE.

  11. Richardson, C., Oster, N., Henriksen, D. & Mishra, P. (2023). “Artificial Intelligence, Responsible Innovation, and the Future of Humanity with Andrew Maynard.” TechTrends. doi:10.1007/s11528-023-00921-2. An author copy is at punyamishra.com/wp-content/uploads/2023/12/TechTrends-Maynard.pdf.

    • What it is: a long interview in a journal, setting out his AI and responsible-innovation views as of late 2023.
    • Corpus coverage: NONE.

  12. Selected 2020science.org archive posts. The blog ran from 2007 to 2019 and has 526 posts; none from before 2014 are in the corpus. These are free.

    • “What was worrying us about nanotechnology safety seven years ago?” (2011-08-09)
    • “Exploring speculated catastrophe and mundane reality” (2012-02-04)
    • “Beyond the obvious – lessons from the Deepwater Horizon Oil Spill” (2010-10-25)
    • “Ten things everyone should know about nanotechnology safety” (2009-08-29)
    • “Navigating the nanotechnology risk landscape – pointers for early career scientists” (2016-07-28)
    • “Is novelty in nanomaterials overrated when it comes to risk?” (2015-02-02)
    • The 2008 “Late lessons” post is listed under item 2.

LOW priority#

Add these only if a specific thread needs them.

No action needed: already covered or already held#

(b) Full relevant bibliography, with coverage status#

Each entry gives the citation, a DOI or URL, the open-access (OA) status, a one-line description and the corpus coverage. Coverage labels are defined at the top of this file. Where a post is named, its slug in working/maynard/corpus/ is given with the date.

B1. Books#

B2. AI-focused preprints, 2026#

B3. Risk innovation, responsible innovation and technology transitions (journals and chapters), 2012–2026#

B4. Nature Nanotechnology commentaries and “Thesis” columns, 2014–2016#

All of these are closed access unless marked otherwise.

B5. Landmark earlier risk and governance items (pre-2014)#

B6. Reports and white papers#

B7. Long-form essays published outside the Substack#

B8. Out of scope but by Maynard (listed for completeness)#

Disambiguation. The correct identifiers are: - ORCID 0000-0003-2117-5128 (“Andrew David Maynard”; PhD Cambridge 1992; UK HSE, then NIOSH, then the Woodrow Wilson Center PEN, then the University of Michigan Risk Science Center, then ASU; moving to the ASU Thunderbird School of Global Management in August 2026). - Google Scholar b8NhWc4AAAAJ. - OpenAlex A5090218914.

The OpenAlex record for A5090218914 wrongly merges works by other people: - Ethics and Social Welfare editorials, Values and Ethics in Social Work and related chapters, and an undergraduate-dissertation study. These are by Andrew Maynard, a UK social-work academic at UEA and Anglia Ruskin. - Aedes albopictus mosquito genetics papers. These belong to Andrew J. Maynard (Australia). - Necrotizing enterocolitis, cystic-fibrosis, angioedema and NAFLD papers, and the MGAT1 cancer abstract. These are medical authors named Andrew Maynard. - Plasma-electrode Pockels cell, asphalt certification, Tasmanian irrigation, a 1934 BMJ letter, a 1977 trout bioassay, and Spanish physician-supply papers.

Crossref author searches also return other people with the surname: - Alan Maynard (health economist) - Ashley Maynard (psychologist) - Alex Maynard (econometrician) - Andrew T. Maynard (GSK medicinal chemist) - Andreza Maynard (Brazilian historian) - Amy Maynard (comics scholar) - an Andrew Maynard at NASA working on propulsion

The DOI prefix 10.64628/… belongs to The Conversation articles that OpenAlex and Crossref index as “other/posted-content”. Only a subset of his Conversation articles are indexed that way.

Gaps and caveats. - Modem Futura podcast episodes were excluded, as the user asked. Other audio and video (Mission: Interplanetary, Future Out Loud, Risk Bites, media interviews) was also left out. - The Google Scholar entry “Constitutive resonance: AI, the transformation of self, and the narrative structures that reveal what theory cannot” (2026) could not be found anywhere online. Andrew should supply it if it exists as a separate paper. - Paywalled items were not read. For the Nature Nanotechnology and Nature columns, coverage judgements rest on titles and Nature’s standfirsts plus corpus greps. The substance of each column is inferred, not verified. - SSRN was not queried directly (it blocks automated access). SSRN items were identified through Crossref DOIs, the CV and andrewmaynard.net, so a very recent SSRN-only posting could be missing. - Medium archive. Maynard reports 100 or more Medium posts since 2014; the corpus holds about 80 posts from 2014–2022. Pre-2023 Medium posts that were never imported into Substack may be missing, and these were not enumerated. - 2020science.org archive. It was sampled by keyword search (“late lessons”, “early warnings”, “precaution”, “risk innovation”, “artificial intelligence”). The whole archive was not reviewed. - Slate Future Tense, WEF Agenda and similar outlets were checked only against the CV’s “selected” list. They were not searched comprehensively. - EEA chapter pages. The page range of the 2013 chapter comes from Maynard’s CV (pp. 562–591). The EEA report text was not opened, to respect the strand separation in CLAUDE.md. The PDF page offset (report page = PDF page − 2) comes from CLAUDE.md. - Scopus and Web of Science were not available. Books and chapters without DOIs (for example the 2017 VVEV essay and the Jurimetrics review) came from the CV and from web search.