Pith. sign in

REVIEW 3 major objections 5 minor 1 cited by

Towards Responsible Governing AI Proliferation

T0 review · 3 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read AI governance built on compute thresholds is being overtaken by a Proliferation paradigm of smaller, hidden, augmented, decentralized, and open-weight models.

desk verdict A careful, well-cited synthesis that usefully names a real governance shift, but its keystone is a fragile efficiency extrapolation that the paper itself flags and then leans on. read the letter →

arxiv 2412.13821 v1 pith:PL2FJXRR submitted 2024-12-18 cs.CY

Edward Kembery This is my paper · ORCID
classification cs.CY
keywords AIgovernanceBigComputeparadigmProliferationSHADOWframeworkthresholdsopen-weightmodelsdecentralizedinformationsecurity
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper argues that current AI risk governance rests on a 'Big Compute' paradigm—the assumption that dangerous capabilities require immense, trackable, expensive compute concentrated in a few frontier labs. It contends that five converging technology pathways (small, hidden, augmented, decentralized, and open-weight models) are creating a 'Proliferation' paradigm in which these assumptions no longer hold. If the paper is right, compute thresholds will become less reliable tripwires, and governance must expand to cover algorithms, decentralized compute, and dangerous information. The paper proposes responsible access policies, privacy-preserving oversight, and information-security measures, while acknowledging that each has serious limitations and may fail in a 'vulnerable world'.

What carries the argument

The argument is carried by the 'SHADOW' framework, a named map of five emerging technology pathways—small, hidden, augmented, decentralized, and open-weight models—that together define the Proliferation paradigm. Working alongside it is the 'AI triad' (algorithms, compute, and informational inputs), which reframes where governance can intervene when compute is no longer a reliable proxy. The 'Big Compute' paradigm is the baseline being challenged: it treats compute as detectable, excludable, quantifiable, and concentrated, which makes governance mechanisms like compute thresholds, responsible scaling policies, and Know-Your-Customer schemes feasible. The paper uses these objects to show how each SHADOW pathway attacks a different assumption in that baseline.

What would settle it

Track the cost to train a model that reaches a fixed, dangerous benchmark over the next five years. If that cost does not fall by the projected factor of roughly 1,000, and dangerous capabilities remain concentrated in models above current compute thresholds, the Proliferation paradigm's core pathway fails.

Watch

Extended reading notes

Core claim

The central claim is that contemporary AI technologies are rapidly diverging from the assumptions that underpin compute-based governance. The paper names the emerging alternative the 'Proliferation' paradigm: a developing network of smaller, decentralized, open-sourced models that are easier to augment, easier to train undetected, and harder for regulators to see, control, or reverse once released. It introduces the 'SHADOW' framework to map five pathways—small models, hidden models, augmented models, decentralized processes, and open-weight models—and argues these interoperate to erode the visibility, enforceability, and reversibility that made compute thresholds attractive. The paper concludes that responsible governance in this paradigm must target all three elements of the 'AI triad'—algorithms, compute, and information—using structured access, privacy-preserving oversight, and careful information security, and that each strategy depends on empirical estimates of the net uplift in malicious versus benevolent capabilities.

Load-bearing premise

The argument depends on algorithmic efficiency continuing to improve at recent historical rates, so that dangerous AI capabilities keep becoming accessible with dramatically less compute.

Editorial extensions

If this is right

  • Compute thresholds will lose predictive power as small models approach frontier capabilities, forcing evaluations to focus on capabilities rather than training compute.
  • Governance must add the other two legs of the AI triad—algorithms and dangerous information—to its toolbox, not just compute.
  • Decentralized compute networks and open-weight releases make harms harder to reverse, so decisions to fund or publish these technologies should be weighed against irreversible-risk thresholds.
  • Responsible access policies, privacy-preserving oversight, and information-security regimes only work if backed by empirical estimates of net capability uplift for both malicious and benevolent actors.
  • The paper's 'accelerate when reversible, slow or pause when irreversible' principle offers a practical heuristic for calibrating all three strategies.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Beyond the paper: if algorithmic efficiency gains continue at the cited rate, the cost of training to a given capability could fall by roughly three orders of magnitude by 2029, which would make compute thresholds nearly meaningless for catching dangerous small models.
  • Beyond the paper: the same logic implies that open-weight releases become the dominant irreversible step, so the most tractable near-term governance lever may be controlling publication of weights and capability keys rather than compute itself.
  • Beyond the paper: the 'vulnerable world' scenario the paper warns about could be tested empirically by tracking whether dangerous capabilities remain confined to large-scale training runs or begin appearing in models trained on consumer hardware.
  • Beyond the paper: a useful extension would be a quantitative model of the access-security trade-off, estimating the marginal uplift in malicious and benevolent actor capabilities as access to weights, fine-tuning, and compute increases.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper argues that existing AI governance mechanisms are built on a 'Big Compute' paradigm—the assumption that frontier AI capabilities require massive, trackable computational infrastructure—and that this paradigm is being undermined by five interoperating trends: Small models, Hidden models, Augmented models, Decentralized processes, and Open-Weight models ('SHADOW'). It introduces the 'Proliferation' paradigm as a future in which dangerous capabilities are more widely diffused, less visible to regulators, and harder to govern. The paper then proposes governance strategies organized around structured access to algorithms, privacy-preserving oversight of decentralized compute, and information-security policies for capability keys and model weights, while emphasizing the need for empirical research on net marginal capability uplifts for both benevolent and malicious actors.

Significance. If accepted, the Proliferation paradigm would reorient AI governance away from compute thresholds toward algorithmic access, decentralized infrastructure, and information control. The paper's main strengths are its broad and careful synthesis of current technical developments; its balanced treatment of benefits, risks, and ethical trade-offs; and its detailed articulation of open research questions. The SHADOW framework is a genuinely useful descriptive device, and the discussion of infohazard-style policies for model weights and jailbreaks is valuable. The manuscript is, however, a conceptual and agenda-setting piece rather than an empirical study: it contains no original quantitative evidence, and its central probabilistic claim rests on extrapolations that are acknowledged but not critically interrogated.

major comments (3)
  1. [Abstract; §3.1] The central claim that the Proliferation paradigm is 'probable' rests on the extrapolation in §3.1 that algorithmic efficiency will continue to double every 8–16 months and that the compute needed for 'any given capability' will fall by roughly a factor of 1000 by 2029. The paper itself concedes that 'unknown ceilings may cause progress to plateau', but this caveat is not propagated into the abstract's 'probable' or into the framing of the paradigm as a near-term shift (§3, §5). The 95% confidence interval of 5.3–13 months cited from Ho et al. already implies a 2.5x range in doubling time, and the benchmark-based efficiency trends (ImageNet, perplexity) are not shown to transfer to high-impact dangerous capabilities such as cyber-exploitation or biological design. Because this projection is load-bearing for the paper's central claim, the manuscript should either soften the modal claim (e.g., to 'plausible' or 'a scenario to prepare for') or provide a sensitivity analysis showing how the Proliferation paradigm fares under slower efficiency growth.
  2. [§3.1 and §3.3] The argument moves from examples of small models that perform well on broad benchmarks (Phi-3, OpenELM, DBRX) and from augmentation techniques (prompting, fine-tuning) to the conclusion that 'dangerously powerful small models' are likely; however, no example of a small model demonstrating a dangerous capability relevant to governance (e.g., bioweapon design, autonomous cyberattack) is provided. This is a load-bearing gap because the risk-mitigation case depends on the capability route, not on parameter-count reduction alone. The manuscript should explicitly state that this route is currently hypothetical and distinguish benchmark-level capability from task-level dangerous capability, with a discussion of why efficiency gains on benchmarks might or might not transfer.
  3. [§3.4 and §4.2] The decentralized compute pathway is presented as one of five SHADOW pathways, but the paper's own evidence indicates that decentralized networks are currently very small (85 A100 GPUs on Akash versus roughly 5,400 for a mid-tier lab) and that the distributed training approach (DiPaCo) is still theoretical. The governance recommendations in §4.2 (thresholds for anonymous use, workload monitoring, potential shutdown) are conditional on this pathway maturing, yet the paper does not state the conditionality or give a time horizon. The authors should mark these proposals as scenario-contingent rather than near-term policy options.
minor comments (5)
  1. [Front matter] The note at the top states that the text is a lightly edited MPhil dissertation from July 2024 and directs readers to consult more recent publications; for a journal submission this provenance note should be removed and the manuscript updated or clearly dated as a preprint.
  2. [§2.2] The phrase 'offer an supplementary paradigm' should read 'offer a supplementary paradigm'.
  3. [§3.1, §4.2] Several cross-references are inconsistent: §3.1 refers to 'section 2.3', '2.4', and '2.5' for pathways that appear in §§3.3–3.5, and §4.2 refers to 'Section 1.2' for the discussion of Big Compute governance in §2.2. These should be corrected.
  4. [Figure 1] The informal statement 'It took about 2 minutes to find a HuggingFace post...' is not reproducible and the screenshot is not described in enough detail; if the figure is retained, it should include a date and a stable citation, or the claim should be moved to a footnote with a link.
  5. [§4.1] The sentence beginning 'Second, policymakers need a more general view of the net marginal uplift...' is repeated almost verbatim in the following paragraph; this should be consolidated for clarity.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the Proliferation/SHADOW framework is a conceptual synthesis grounded in external empirical evidence, not in fitted parameters or self-citation.

full rationale

The paper makes no formal derivation and does not fit parameters to data; its central claim is a conceptual and predictive argument that contemporary AI developments are moving away from the 'Big Compute' assumptions. The load-bearing evidence is external and non-circular: efficiency-doubling estimates are attributed to Hernandez and Brown and to Ho et al.; the 1000x cost-reduction projection is attributed to the CNAS report; specific model and infrastructure examples (Phi-3, OpenELM, DBRX, DiPaCo, Akash) are cited from independent authors. The SHADOW categories are introduced as a mapping tool, not as a derivation from their own definitions: the paper explicitly notes that 'SHADOW' is 'a tool for beginning a conversation' and that the paradigm is 'necessarily uncertain.' The definition of 'small models' as below compute thresholds is stipulative, but the paper's claims that such models are becoming technically viable and more widely accessible are supported by external sources and are not equivalent to the definition. The paper even flags the fragility of its own extrapolation in Section 3.1: 'Timelines are unclear... unknown ceilings may cause progress to plateau.' That is an acknowledged limitation of an empirical forecast, not a circular step. No parameter fitted to a target is later renamed as a prediction, and no load-bearing conclusion rests solely on a self-citation. The paper is therefore self-contained as a governance analysis and has no significant circularity.

Assumptions & free parameters 0 free parameters · 4 assumptions · 3 invented entities

The paper's argument rests on several domain assumptions from AI scaling laws and efficiency trends, plus the philosophical framing of technological paradigms. No numbers are fitted in this paper.

assumptions (4)
  • domain assumption AI capabilities scale with compute, so compute is a useful proxy for risk.
    Invoked in Section 2.2 to describe Big Compute; the paper treats it as an empirical regularity cited from prior work.
  • domain assumption Algorithmic efficiency gains will continue at historical rates, making smaller models more capable.
    Section 3.1 relies on trends from Brown and Hernandez and Ho et al. to project that dangerous capabilities will become accessible with less compute.
  • domain assumption The risk-based case for AI governance is valid; severe harms from AI are sufficiently plausible to warrant governance.
    Section 2.1 lists potential harms but does not quantify their probability; the paper assumes these risks justify regulation.
  • domain assumption Dosi's technological paradigm framework is a useful lens for understanding AI development.
    Section 2.2 applies this philosophy-of-science framework; it is not empirically tested but shapes the paper's structure.
invented entities (3)
  • Proliferation paradigm
    purpose: To describe the shift to small, hidden, augmented, decentralized, and open-weight AI.
    The paper's own conceptual umbrella; not directly measurable.
  • SHADOW framework
    purpose: To organize five pathways that define the paradigm.
    A taxonomy generated by the author to structure the discussion.
  • Capability keys independent evidence
    purpose: To name information (prompts, fine-tuning data, architectures) that can elicit dangerous capabilities without retraining.
    The existence of jailbreaks and open weights provides evidence outside the paper; the term is a metaphor.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Towards Responsible Governing AI Proliferation." pith.science (2026). https://pith.science/paper/PL2FJXRR

@misc{pith2026241213821,
  author       = {Pith},
  title        = {Pith review of: Towards Responsible Governing AI Proliferation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PL2FJXRR}},
  note         = {Machine review of arXiv:2412.13821}
}
read the original abstract

This paper argues that existing governance mechanisms for mitigating risks from AI systems are based on the `Big Compute' paradigm -- a set of assumptions about the relationship between AI capabilities and infrastructure -- that may not hold in the future. To address this, the paper introduces the `Proliferation' paradigm, which anticipates the rise of smaller, decentralized, open-sourced AI models which are easier to augment, and easier to train without being detected. It posits that these developments are both probable and likely to introduce both benefits and novel risks that are difficult to mitigate through existing governance mechanisms. The final section explores governance strategies to address these risks, focusing on access governance, decentralized compute oversight, and information security. Whilst these strategies offer potential solutions, the paper acknowledges their limitations and cautions developers to weigh benefits against developments that could lead to a `vulnerable world'.

Figures

Figures reproduced from arXiv: 2412.13821 by the authors.

Figure 1
Figure 1. It took about 2 minutes to find a HuggingFace post describing how [PITH_FULL_IMAGE:figures/full_fig_p027_1.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. LLM Harms: A Taxonomy and Discussion

    cs.CY 2025-12 unverdicted novelty 3.0 of 10

    This paper proposes a taxonomy of LLM harms in five categories and suggests mitigation strategies plus a dynamic auditing system for responsible development.

Reference graph

Works this paper leans on

153 extracted references · 33 canonical work pages · cited by 1 Pith paper

  1. [1]

    The Making of the Atomic Bomb

    Richard Rhodes. The Making of the Atomic Bomb . Simon and Schuster, 2012

    2012

  2. [2]

    The vulnerable world hypothesis

    Nick Bostrom. “The vulnerable world hypothesis”. In: Global Policy 10.4 (2019), pp. 455–476

    2019

  3. [3]

    Compute trends across three eras of machine learn- ing

    Jaime Sevilla et al. “Compute trends across three eras of machine learn- ing”. In: 2022 International Joint Conference on Neural Networks (IJCNN). IEEE. 2022, pp. 1–8

    2022

  4. [4]

    Computing Power and the Governance of Artificial Intelligence

    Girish Sastry et al. “Computing Power and the Governance of Artificial Intelligence”. In: arXiv preprint arXiv:2402.08797 (2024)

    arXiv 2024

  5. [5]

    Frontier AI regulation: Managing emerging risks to public safety

    Markus Anderljung et al. “Frontier AI regulation: Managing emerging risks to public safety”. In: arXiv preprint arXiv:2307.03718 (2023)

    arXiv 2023

  6. [6]

    Training Compute Thresholds: Features and Functions in AI Governance

    Lennart Heim. “Training Compute Thresholds: Features and Functions in AI Governance”. In: arXiv preprint arXiv:2405.10799 (2024)

    arXiv 2024

  7. [7]

    Technological paradigms and technological trajectories: a suggested interpretation of the determinants and directions of technical change

    Giovanni Dosi. “Technological paradigms and technological trajectories: a suggested interpretation of the determinants and directions of technical change”. In: Research policy 11.3 (1982), pp. 147–162

    1982

  8. [8]

    Anthropic’s Responsible Scaling Policy

    Anthropic. Anthropic’s Responsible Scaling Policy. Tech. rep. Anthropic, 2023

    2023

Show all 153 references
  1. [9]

    Introducing the Frontier Safety Framework

    Anca Dragan, Helen King, and Allan Dafoe. Introducing the Frontier Safety Framework. Tech. rep. Google DeepMind, 2024

    2024

  2. [10]

    Preparedness Framework (Beta)

    OpenAI. Preparedness Framework (Beta). Tech. rep. OpenAI, December 18 2023

    2023

  3. [12]

    AI Governance

    Allan Dafoe. “AI Governance”. In: The Oxford Handbook of AI Gover- nance (2024), p. 21

    2024

  4. [13]

    AI governance for businesses

    Johannes Schneider et al. “AI governance for businesses”. In: arXiv preprint arXiv:2011.10672 (2020)

    2020 arXiv

  5. [14]

    Governance of artificial intelligence

    Araz Taeihagh. “Governance of artificial intelligence”. In: Policy and so- ciety 40.2 (2021), pp. 137–157

    2021

  6. [15]

    International institutions for advanced AI

    Lewis Ho et al. “International institutions for advanced AI”. In: arXiv preprint arXiv:2307.04699 (2023)

    2023 arXiv

  7. [16]

    Model evaluation for extreme risks

    Toby Shevlane et al. “Model evaluation for extreme risks”. In: arXiv preprint arXiv:2305.15324 (2023)

    2023 arXiv

  8. [17]

    The Malicious Use of Artificial Intelligence: Fore- casting, Prevention, and Mitigation

    Miles Brundage et al. The Malicious Use of Artificial Intelligence: Fore- casting, Prevention, and Mitigation . 2018. arXiv: 1802.07228 [cs.AI]

    2018 arXiv

  9. [18]

    Artificial intelligence and biological misuse: Differ- entiating risks of language models and biological design tools

    Jonas B Sandbrink. “Artificial intelligence and biological misuse: Differ- entiating risks of language models and biological design tools”. In: arXiv preprint arXiv:2306.13952 (2023). 30

    2023 arXiv

  10. [19]

    Building an early warning system for LLM-aided biological threat creation

    OpenAI. Building an early warning system for LLM-aided biological threat creation. Tech. rep. OpenAI, 2024)

    2024

  11. [20]

    The threat of offensive ai to organizations

    Yisroel Mirsky et al. “The threat of offensive ai to organizations”. In: Computers and Security 124 (2023), p. 103006

    2023

  12. [21]

    Artificial intelligence (AI) cybersecurity dimensions: a comprehensive framework for understanding adversarial and offensive AI

    Masike Malatji and Alaa Tolah. “Artificial intelligence (AI) cybersecurity dimensions: a comprehensive framework for understanding adversarial and offensive AI”. In: AI and Ethics (2024), pp. 1–28

    2024

  13. [22]

    Ai alignment: A comprehensive survey

    Jiaming Ji et al. “Ai alignment: A comprehensive survey”. In: arXiv preprint arXiv:2310.19852 (2023)

    2023 arXiv

  14. [23]

    Foundational challenges in assuring alignment and safety of large language models

    Usman Anwar et al. “Foundational challenges in assuring alignment and safety of large language models”. In: arXiv preprint arXiv:2404.09932 (2024)

    2024 arXiv

  15. [24]

    Harms of AI

    Daron Acemoglu. Harms of AI . Tech. rep. National Bureau of Economic Research, 2021

    2021

  16. [25]

    The incompatible incentives of private-sector AI

    Tom Slee. “The incompatible incentives of private-sector AI”. In: The Oxford Handbook of Ethics of AI (2020), pp. 106–123

    2020

  17. [26]

    Towards Responsible Governance of Biological Design Tools

    Richard Moulange et al. Towards Responsible Governance of Biological Design Tools. 2023. arXiv: 2311.15936 [cs.CY]

    2023 arXiv

  18. [27]

    A framework for fairness: A systematic review of existing fair ai solutions

    Brianna Richardson and Juan E Gilbert. “A framework for fairness: A systematic review of existing fair ai solutions”. In: arXiv preprint arXiv:2112.05700 (2021)

    2021 arXiv

  19. [28]

    Data, Privacy, and the Individual: A Report for the Cen- ter for the Governance of Change (https://philarchive.org/rec/VLIPM)

    Carissa V´ eliz. “Data, Privacy, and the Individual: A Report for the Cen- ter for the Governance of Change (https://philarchive.org/rec/VLIPM)”. In: (2020)

    2020

  20. [29]

    Privacy is Power

    Carissa V´ eliz. Privacy is Power . London, UK: Penguin (Bantam Press), 2020

    2020

  21. [30]

    Account- ability in artificial intelligence: what it is and how it works

    Claudio Novelli, Mariarosaria Taddeo, and Luciano Floridi. “Account- ability in artificial intelligence: what it is and how it works”. In: AI and SOCIETY (2023), pp. 1–12

    2023

  22. [31]

    Democratising AI: Multiple meanings, goals, and methods

    Elizabeth Seger et al. “Democratising AI: Multiple meanings, goals, and methods”. In: Proceedings of the 2023 AAAI/ACM Conference on AI, Ethics, and Society . 2023, pp. 715–722

    2023

  23. [32]

    The case for strong longter- mism

    Hilary Greaves and William MacAskill. “The case for strong longter- mism”. In: Global Priorities Institute Working Paper No. 5-2021 (2021)

    2021

  24. [33]

    Global AI governance: barriers and pathways for- ward

    Huw Roberts et al. “Global AI governance: barriers and pathways for- ward”. In: International Affairs 100.3 (2024), pp. 1275–1286

    2024

  25. [34]

    AI, governance and ethics: global perspectives

    Angela Daly et al. “AI, governance and ethics: global perspectives”. In: University of Hong Kong Faculty of Law Research Paper 2020/051 (2020)

    2020

  26. [35]

    Artificial intelligence and China’s authoritarian gover- nance

    Jinghan Zeng. “Artificial intelligence and China’s authoritarian gover- nance”. In: International Affairs 96.6 (2020), pp. 1441–1459. 31

    2020

  27. [36]

    AI governance: a research agenda

    Allan Dafoe. “AI governance: a research agenda”. In: Governance of AI Program, Future of Humanity Institute, University of Oxford: Oxford, UK 1442 (2018), p. 1443

    2018

  28. [37]

    Inductive Risk and Values in Science

    Heather Douglas. “Inductive Risk and Values in Science”. In: Philosophy of Science 67.4 (2000), pp. 559–579. issn: 00318248, 1539767X. url: http://www.jstor.org/stable/188707 (visited on 06/04/2024)

    2000

  29. [38]

    A framework of blockchain technology adoption: An investigation of challenges and ex- pected value

    Elissar Toufaily, Tatiana Zalan, and Soumaya Ben Dhaou. “A framework of blockchain technology adoption: An investigation of challenges and ex- pected value”. In: Information and Management 58.3 (2021), p. 103444

    2021

  30. [39]

    Digital technologies, innovation, and skills: Emerg- ing trajectories and challenges

    Tommaso Ciarli et al. “Digital technologies, innovation, and skills: Emerg- ing trajectories and challenges”. In:Research Policy50.7 (2021), p. 104289

    2021

  31. [40]

    A typology of technological change: Technological paradigm theory with validation and generalization from case studies

    Jonathan C Ho and Chung-Shing Lee. “A typology of technological change: Technological paradigm theory with validation and generalization from case studies”. In: Technological Forecasting and Social Change97 (2015), pp. 128–139

    2015

  32. [41]

    ”The AI Triad and What It Means for National Secu- rity Strategy”

    Ben Buchanan. ”The AI Triad and What It Means for National Secu- rity Strategy”. Tech. rep. Center for Security and Emerging Technology, August 2020

    2020

  33. [42]

    Training compute-optimal large language mod- els

    Jordan Hoffmann et al. “Training compute-optimal large language mod- els”. In: arXiv preprint arXiv:2203.15556 (2022)

    2022 arXiv

  34. [43]

    How predictable is language model benchmark performance?

    David Owen. How predictable is language model benchmark performance?

  35. [44]

    Racing to the Trillion-Dollar Cluster

    Leopold Aschenbrenner. Racing to the Trillion-Dollar Cluster . Tech. rep. Situational Awareness (Blog), June 6th 2024

    2024

  36. [45]

    Study of Cloud Security in Hyper- scalers

    Megha Panjwani and Suman De. “Study of Cloud Security in Hyper- scalers”. In: (2020). doi: https://doi.org/10.23919/indiacom49435. 2020.9083727

    2020

  37. [46]

    Ethical issues in the big data industry

    Kirsten E Martin. “Ethical issues in the big data industry”. In: Strategic Information Management. Routledge, 2020, pp. 450–471

    2020

  38. [47]

    ’An Ecosystem Framework of AI Governance’

    Bernd W. Wirtz, Paul F. Langer, and Jan C. Weyerer. “’An Ecosystem Framework of AI Governance’”. In: The Oxford Handbook of AI Gover- nance, Justin B. Bullock, and others (eds) (2024)

    2024

  39. [48]

    Oversight for Frontier AI through a Know-Your-Customer Scheme for Compute Providers

    Janet Egan and Lennart Heim. Oversight for Frontier AI through a Know-Your-Customer Scheme for Compute Providers. 2023. arXiv: 2310. 13625 [cs.CY]

    2023

  40. [49]

    Executive order on the safe, secure, and trustworthy development and use of artificial intelligence

    Joseph R Biden. “Executive order on the safe, secure, and trustworthy development and use of artificial intelligence”. In: (2023)

    2023

  41. [50]

    The structure of scientific revolutions

    Thomas S Kuhn. The structure of scientific revolutions . Vol. 962. Uni- versity of Chicago press Chicago, 1997

    1997

  42. [51]

    Future-Proofing Frontier AI Regulation

    Paul Scharre. Future-Proofing Frontier AI Regulation. Tech. rep. Center For National American Security, 2024. 32

    2024

  43. [52]

    Danny Hernandez and Tom B. Brown. Measuring the Algorithmic Effi- ciency of Neural Networks . 2020. arXiv: 2005.04305 [cs.LG]

    2020 arXiv

  44. [53]

    Algorithmic progress in language models

    Anson Ho et al. Algorithmic progress in language models . 2024. arXiv: 2403.05812 [cs.CL]

    2024 arXiv

  45. [54]

    A Survey of Quantization Methods for Efficient Neural Network Inference

    Amir Gholami et al. A Survey of Quantization Methods for Efficient Neural Network Inference. 2021. arXiv: 2103.13630 [cs.CV]

    2021 arXiv

  46. [55]

    A Survey on Deep Neural Network Pruning-Taxonomy, Comparison, Analysis, and Recom- mendations

    Hongrong Cheng, Miao Zhang, and Javen Qinfeng Shi. A Survey on Deep Neural Network Pruning-Taxonomy, Comparison, Analysis, and Recom- mendations. 2023. arXiv: 2308.06767 [cs.LG]

    2023 arXiv

  47. [56]

    A Comprehensive Review of Knowledge Dis- tillation in Computer Vision

    Sheikh Musa Kaleem et al. A Comprehensive Review of Knowledge Dis- tillation in Computer Vision . 2024. arXiv: 2404.00936 [cs.CV]

    2024 arXiv

  48. [57]

    Parameter-Efficient Fine-Tuning Methods for Pre- trained Language Models: A Critical Review and Assessment

    Lingling Xu et al. Parameter-Efficient Fine-Tuning Methods for Pre- trained Language Models: A Critical Review and Assessment. 2023. arXiv: 2312.12148 [cs.CL]

    2023 arXiv

  49. [58]

    Phi-3 technical report: A highly capable language model locally on your phone

    Marah Abdin et al. “Phi-3 technical report: A highly capable language model locally on your phone”. In:arXiv preprint arXiv:2404.14219 (2024)

    2024 arXiv

  50. [59]

    Brown et al

    Tom B. Brown et al. Language Models are Few-Shot Learners . 2020. arXiv: 2005.14165 [cs.CL]

    2020 arXiv

  51. [60]

    OpenELM: An Efficient Language Model Family with Open Training and Inference Framework

    Sachin Mehta et al. OpenELM: An Efficient Language Model Family with Open Training and Inference Framework. 2024. arXiv: 2404.14619 [cs.CL]

    2024 arXiv

  52. [61]

    Outrageously Large Neural Networks: The Sparsely- Gated Mixture-of-Experts Layer

    Noam Shazeer et al. Outrageously Large Neural Networks: The Sparsely- Gated Mixture-of-Experts Layer. 2017. arXiv: 1701.06538 [cs.LG]

    2017 arXiv

  53. [62]

    Introducing DBRX: A New State-of-the-Art Open LLM

    Mosaic Research Team. Introducing DBRX: A New State-of-the-Art Open LLM. Tech. rep. Databricks, March 27, 2024)

    2024

  54. [63]

    Snowflake Arctic: The Best LLM for Enterprise AI — Efficiently Intelligent, Truly Open

    Snowflake AI Research. Snowflake Arctic: The Best LLM for Enterprise AI — Efficiently Intelligent, Truly Open . Tech. rep. Snowflake, April 24 2024

    2024

  55. [64]

    The Bitter-er Lesson

    Aidan McLaughlin. The Bitter-er Lesson . Tech. rep. Personal Blog, 14 June 2024

    2024

  56. [65]

    KAN: Kolmogorov-Arnold Networks

    Ziming Liu et al. KAN: Kolmogorov-Arnold Networks. 2024. arXiv: 2404. 19756 [cs.LG]

    2024

  57. [66]

    Position: Leverage Foundational Models for Black- Box Optimization

    Xingyou Song et al. Position: Leverage Foundational Models for Black- Box Optimization. 2024. arXiv: 2405.03547

    2024 arXiv

  58. [68]

    MobiLlama: Towards Accurate and Lightweight Fully Transparent GPT

    Omkar Thawakar et al. “MobiLlama: Towards Accurate and Lightweight Fully Transparent GPT”. In: arXiv preprint arXiv:2402.16840 (2024)

    2024 arXiv

  59. [69]

    Phi 3 and Arctic: Outlier LMs are hints

    Nathan Lambert. Phi 3 and Arctic: Outlier LMs are hints . Tech. rep. Interconnects.ai, April 30 2024. 33

    2024

  60. [70]

    Increased Compute Efficiency and the Diffusion of AI Capabilities

    Konstantin Pilz, Lennart Heim, and Nicholas Brown. Increased Compute Efficiency and the Diffusion of AI Capabilities . 2024. arXiv: 2311.15377 [cs.CY]

    2024 arXiv

  61. [71]

    GPT-2: The Strange Debacle Surrounding Mystery ‘OpenAI’ Chatbot

    Tim Keary. GPT-2: The Strange Debacle Surrounding Mystery ‘OpenAI’ Chatbot. Tech. rep. Techopedia, 11 May 2024

    2024

  62. [72]

    Responsible Reporting for Frontier AI Development

    Noam Kolt et al. Responsible Reporting for Frontier AI Development

  63. [73]

    arXiv: 2404.02675 [cs.CY]

    arXiv

  64. [74]

    Qwen technical report

    Jinze Bai et al. “Qwen technical report”. In: arXiv preprint arXiv:2309.16609 (2023)

    2023 arXiv

  65. [75]

    Through partnership with IBM, Saudi Data and Artificial Intelli- gence Authority (SDAIA) launches a groundbreaking Arabic AI model to the Middle East

    IBM. Through partnership with IBM, Saudi Data and Artificial Intelli- gence Authority (SDAIA) launches a groundbreaking Arabic AI model to the Middle East . Tech. rep. IBM, May 21 2024

    2024

  66. [76]

    Abu Dhabi makes its Falcon 40B AI model open sourc

    Lisa Barrington. Abu Dhabi makes its Falcon 40B AI model open sourc . Tech. rep. Reuters, May 25 2023

    2023

  67. [77]

    Near to Mid-term Risks and Opportunities of Open Source Generative AI

    Francisco Eiras et al. “Near to Mid-term Risks and Opportunities of Open Source Generative AI”. In:arXiv preprint arXiv:2404.17047 (2024)

    2024 arXiv

  68. [78]

    Chain-of-Thought Prompting Elicits Reasoning in Large Language Models

    Jason Wei et al. Chain-of-Thought Prompting Elicits Reasoning in Large Language Models. 2023. arXiv: 2201.11903 [cs.CL]

    2023 arXiv

  69. [79]

    Meta-Prompting: Enhancing Language Models with Task-Agnostic Scaffolding

    Mirac Suzgun and Adam Tauman Kalai. Meta-Prompting: Enhancing Language Models with Task-Agnostic Scaffolding . 2024. arXiv: 2401 . 12954 [cs.CL]

    2024

  70. [80]

    AI capabilities can be significantly improved without expensive retraining

    Tom Davidson et al. AI capabilities can be significantly improved without expensive retraining. 2023. arXiv: 2312.07413 [cs.AI]

    2023 arXiv

  71. [81]

    Sholto Douglas and Trenton Bricken - How to Build and Understand GPT-7’s Mind (https://www.dwarkeshpatel.com/p/sholto-douglas- trenton-bricken)

    Dwarkesh Patel. Sholto Douglas and Trenton Bricken - How to Build and Understand GPT-7’s Mind (https://www.dwarkeshpatel.com/p/sholto-douglas- trenton-bricken). Tech. rep. Dwarkesh Podcast, 28 May 2024

    2024

  72. [82]

    Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context

    Gemini Team et al. Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context . 2024. arXiv: 2403.05530 [cs.CL]

    2024 arXiv

  73. [83]

    Parameter-Efficient Fine-Tuning for Large Models: A Comprehensive Survey

    Zeyu Han et al. Parameter-Efficient Fine-Tuning for Large Models: A Comprehensive Survey. 2024. arXiv: 2403.14608 [cs.LG]

    2024 arXiv

  74. [84]

    Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To! 2023

    Xiangyu Qi et al. Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To! 2023. arXiv: 2310.03693 [cs.CL]

    2023 arXiv

  75. [85]

    Evolutionary Optimization of Model Merging Recipes

    Takuya Akiba et al. Evolutionary Optimization of Model Merging Recipes

  76. [86]

    arXiv: 2403.13187 [cs.NE]

    arXiv

  77. [87]

    Command R+ (https://docs.cohere.com/docs/command-r-plus)

    Cohere. Command R+ (https://docs.cohere.com/docs/command-r-plus). Tech. rep. Cohere, 23 May 2024

    2024

  78. [88]

    Github Copilot (https://github.com/features/copilot)

    Github. Github Copilot (https://github.com/features/copilot). Tech. rep. Github, october 2021. 34

    2021

  79. [89]

    AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation Framework

    Qingyun Wu et al. “AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation Framework”. In: 2023. arXiv: 2308.08155 [cs.AI]

    2023 arXiv

  80. [90]

    Training Language Model Agents without Modi- fying Language Models

    Shaokun Zhang et al. “Training Language Model Agents without Modi- fying Language Models”. In: ICML’24 (2024)

    2024

  81. [91]

    Frontier Safety Framework

    Deepmind. Frontier Safety Framework. Tech. rep. Google Deepmind, 17 May 2024

    2024

  82. [92]

    Multi-Stage Prompting for Next Best Agent Rec- ommendations in Adaptive Workflows

    Prerna Agarwal et al. “Multi-Stage Prompting for Next Best Agent Rec- ommendations in Adaptive Workflows”. In: Proceedings of the AAAI Conference on Artificial Intelligence. Vol. 38. 21. 2024, pp. 22843–22849

    2024

  83. [93]

    Deep reinforcement learning from human prefer- ences

    Paul Christiano et al. Deep reinforcement learning from human prefer- ences. 2023. arXiv: 1706.03741 [stat.ML]

    2023 arXiv

  84. [94]

    Distributing and Democratizing Institutional Power Through Decentralization

    Amir Fard Bahreini et al. “Distributing and Democratizing Institutional Power Through Decentralization”. In:Building Decentralized Trust: Mul- tidisciplinary Perspectives on the Design of Blockchains and Distributed Ledgers (2021), pp. 95–109

    2021

  85. [95]

    Big AI can centralize decision-making and power, and that’sa problem

    Erik Brynjolfsson and Andrew Ng. “Big AI can centralize decision-making and power, and that’sa problem”. In: Missing links in ai governance 65 (2023)

    2023

  86. [96]

    Jailbreak and Guard Aligned Language Models with Only Few In-Context Demonstrations

    Zeming Wei et al. Jailbreak and Guard Aligned Language Models with Only Few In-Context Demonstrations. 2024. arXiv: 2310.06387 [cs.LG]

    2024 arXiv

  87. [97]

    Distributed, decentralized, and democratized artificial intelligence

    Gabriel Axel Montes and Ben Goertzel. “Distributed, decentralized, and democratized artificial intelligence”. In: Technological Forecasting and Social Change 141 (2019), pp. 354–358

    2019

  88. [98]

    A Review on Building Blocks of Decentralized Artificial Intelligence

    Vid Kersic and Muhamed Turkanovic. A Review on Building Blocks of Decentralized Artificial Intelligence. 2024. arXiv: 2402.02885 [cs.AI]

    2024 arXiv

  89. [99]

    Decentralization of artificial intelligence: Analyzing de- velopments in decentralized learning and distributed AI networks

    Ishan Gupta. “Decentralization of artificial intelligence: Analyzing de- velopments in decentralized learning and distributed AI networks”. In: arXiv preprint arXiv:1603.04467 (2020)

    2020 arXiv

  90. [100]

    Distributed Machine Learning on Akash Network With Ray (https://akash.network/)

    Anil Murty. Distributed Machine Learning on Akash Network With Ray (https://akash.network/). Tech. rep. Akash, January 28 2024

    2024

  91. [101]

    Decentralized Physical Infrastructure Network (De- PIN): Challenges and Opportunities

    Zhibin Lin et al. “Decentralized Physical Infrastructure Network (De- PIN): Challenges and Opportunities”. In:arXiv preprint arXiv:2406.02239 (2024)

    2024 arXiv

  92. [102]

    Golem (https://www.golem.network/)

    Golem. Golem (https://www.golem.network/). Tech. rep. Golem, 2021

    2021

  93. [103]

    Welcome to DeepBrain Chain (https://www.deepbrainchain.org/)

    DeepBrain Chain. Welcome to DeepBrain Chain (https://www.deepbrainchain.org/). Tech. rep. DeepBrain Chain, 2017

    2017

  94. [104]

    Build, Own and Monetise Web3 (https://iex.ec/)

    iExec. Build, Own and Monetise Web3 (https://iex.ec/). Tech. rep. iExec, 2020. 35

    2020

  95. [105]

    Gensyn Litepaper: The hyperscale, cost-efficient compute protocol for the world’s deep learning models(https://docs.gensyn.ai/litepaper)

    gensyn. Gensyn Litepaper: The hyperscale, cost-efficient compute protocol for the world’s deep learning models(https://docs.gensyn.ai/litepaper) . Tech. rep. Gensyn, 2024

    2024

  96. [106]

    Shard: On the decentralized training of foundation models (https://aksh-garg.medium.com/shard-on-the-decentralized-training-of-foundation- models-2fd982176724)

    Aksh Garg. Shard: On the decentralized training of foundation models (https://aksh-garg.medium.com/shard-on-the-decentralized-training-of-foundation- models-2fd982176724). Tech. rep. Medium, May 20 2024

    2024

  97. [107]

    DiPaCo: Distributed Path Composition

    Arthur Douillard et al. “DiPaCo: Distributed Path Composition”. In: arXiv preprint arXiv:2403.10616 (2024)

    2024 arXiv

  98. [108]

    LinguaLinked: A Distributed Large Language Model Inference System for Mobile Devices

    Junchen Zhao et al. LinguaLinked: A Distributed Large Language Model Inference System for Mobile Devices . 2023. arXiv: 2312.00388 [cs.LG]

    2023 arXiv

  99. [109]

    Open-sourcing highly capable foundation models: An evaluation of risks, benefits, and alternative methods for pursuing open-source objectives

    Elizabeth Seger et al. “Open-sourcing highly capable foundation models: An evaluation of risks, benefits, and alternative methods for pursuing open-source objectives”. In: arXiv preprint arXiv:2311.09227 (2023)

    2023 arXiv

  100. [110]

    Chatbot Arena: An Open Platform for Evaluating LLMs by Human Preference

    Wei-Lin Chiang et al. Chatbot Arena: An Open Platform for Evaluating LLMs by Human Preference . 2024. arXiv: 2403.04132 [cs.AI]

    2024 arXiv

  101. [111]

    Introducing Meta Llama 3: The most capable openly available LLM to date (https://ai.meta.com/blog/meta-llama-3/)

    Meta AI Research. Introducing Meta Llama 3: The most capable openly available LLM to date (https://ai.meta.com/blog/meta-llama-3/) . Tech. rep. Meta, 18 April 2024

    2024

  102. [112]

    Mark Zuckerberg - Llama 3, Open Sourcing 10b Models, Caesar Augustus

    Dwarkesh Patel. Mark Zuckerberg - Llama 3, Open Sourcing 10b Models, Caesar Augustus. Tech. rep. Dwarkesh podcast, 18 April 2024

    2024

  103. [113]

    HugginFace index (https://huggingface.co/docs/hub/en/index)

    HuggingFace. HugginFace index (https://huggingface.co/docs/hub/en/index). Tech. rep. Huggingface, 2024

    2024

  104. [114]

    Open-Source

    David Evan Harris. How to Regulate Unsecured “Open-Source” AI: No Exemptions. Tech. rep. Tech Policy Press, December 4 2023

    2023

  105. [115]

    Open-Source

    David Evan Harris. How to Regulate Unsecured “Open-Source” AI: No Exemptions. Tech. rep. Tech Policy Press, 18 December 2023

    2023

  106. [116]

    Content Policy (https://huggingface.co/content-guidelines)

    HuggingFace. Content Policy (https://huggingface.co/content-guidelines). Tech. rep. HuggingFace, 30 August 2023

    2023

  107. [117]

    Import AI Import AI 375: GPT-2 five years later; decentral- ized training; new ways of thinking about consciousness and AI

    Jack Clark. Import AI Import AI 375: GPT-2 five years later; decentral- ized training; new ways of thinking about consciousness and AI . Tech. rep. Demoscene AI, 2024

    2024

  108. [118]

    Toward a theory of justice for artificial intelligence

    Iason Gabriel. “Toward a theory of justice for artificial intelligence”. In: Daedalus 151.2 (2022), pp. 218–231

    2022

  109. [119]

    How AI Threatens Democracy

    Sarah Kreps and Doug Kriner. “How AI Threatens Democracy”. In: Jour- nal of Democracy 34.4 (2023), pp. 122–131

    2023

  110. [120]

    Artificial intelligence: Risks to privacy and democracy

    Karl Manheim and Lyric Kaplan. “Artificial intelligence: Risks to privacy and democracy”. In: Yale JL and Tech. 21 (2019), p. 106

    2019

  111. [121]

    The role of artificial intelligence in disinformation

    No´ emi Bontridder and Yves Poullet. “The role of artificial intelligence in disinformation”. In: Data and Policy 3 (2021), e32

    2021

  112. [122]

    Internet

    Hannah Ritchie et al. “Internet”. In: Our World in Data (2023). 36

    2023

  113. [123]

    Number of software developers worldwide in 2018 to 2024 (in millions)

    David Evans. “Number of software developers worldwide in 2018 to 2024 (in millions)”. In: Statista ((August 21, 2023))

    2018

  114. [124]

    Maslej N

    et al. Maslej N. Artificial Intelligence Index Report 2023: Chapter 7 Di- versity. Tech. rep. Institute for HumanCentered AI, Stanford University, Stanford, CA, 2023

    2023

  115. [125]

    Artificial intelligence, jobs and the future of work: Racing with the machines

    Edvard PG Bruun and Alban Duka. “Artificial intelligence, jobs and the future of work: Racing with the machines”. In:Basic Income Studies 13.2 (2018), p. 20180018

    2018

  116. [126]

    Structured access: an emerging paradigm for safe AI deployment

    Toby Shevlane. “Structured access: an emerging paradigm for safe AI deployment”. In: arXiv preprint arXiv:2201.05159 (2022)

    2022 arXiv

  117. [127]

    The gradient of generative AI release: Methods and considerations

    Irene Solaiman. “The gradient of generative AI release: Methods and considerations”. In: Proceedings of the 2023 ACM conference on fairness, accountability, and transparency. 2023, pp. 111–122

    2023

  118. [128]

    STRUCTURED ACCESS FOR THIRD-PARTY RESEARCH ON FRONTIER AI MODELS: IN- VESTIGATING RESEARCHERS’MODEL ACCESS REQUIREMENTS

    Benjamin S Bucknall and Robert F Trager. “STRUCTURED ACCESS FOR THIRD-PARTY RESEARCH ON FRONTIER AI MODELS: IN- VESTIGATING RESEARCHERS’MODEL ACCESS REQUIREMENTS”. In: (2023)

    2023

  119. [129]

    Responsible Scaling Policies (https://metr.org/blog/2023-09-26- rsp/)

    METR. Responsible Scaling Policies (https://metr.org/blog/2023-09-26- rsp/). Tech. rep. METR, 26 September 2023

    2023

  120. [130]

    How does the offense-defense balance scale?

    Ben Garfinkel and Allan Dafoe. “How does the offense-defense balance scale?” In: Emerging Technologies and International Stability. Routledge, 2021, pp. 247–274

    2021

  121. [131]

    Societal Adaptation to Advanced AI

    Jamie Bernardi et al. “Societal Adaptation to Advanced AI”. In: arXiv preprint arXiv:2405.10295 (2024)

    2024 arXiv

  122. [132]

    On mapping val- ues in AI governance

    Geoff Gordon, Bernhard Rieder, and Giovanni Sileno. “On mapping val- ues in AI governance”. In:Computer, Law and Security Review 46 (2022), p. 105712

    2022

  123. [133]

    Privacy As Contextual Integrity

    Helen Nissenbaum. “Privacy As Contextual Integrity”. In: Washington Law Review 79 (May 2004)

    2004

  124. [134]

    Market, non-market and anti-market processes in neoliberalism

    Julien Mercille and Enda Murphy. “Market, non-market and anti-market processes in neoliberalism”. In: Critical Sociology 45.7-8 (2019), pp. 1093– 1109

    2019

  125. [135]

    Semiconductor and ICT Industrial Policy in the US and EU: Geopolitical Threat Responses

    Shawn Donnelly. “Semiconductor and ICT Industrial Policy in the US and EU: Geopolitical Threat Responses”. In: Politics and Governance 11.4 (2023), pp. 129–139

    2023

  126. [136]

    Secure, Governable Chips Using On-Chip Mechanisms to Manage National Security Risks from AI and Advanced Computing

    Onni Aarne, Tim Fist, and Caleb Withers. “Secure, Governable Chips Using On-Chip Mechanisms to Manage National Security Risks from AI and Advanced Computing”. In: CNAS (January 2024)

    2024

  127. [137]

    A human rights-based approach to re- sponsible AI

    Vinodkumar Prabhakaran et al. “A human rights-based approach to re- sponsible AI”. In: arXiv preprint arXiv:2210.02667 (2022). 37

    2022 arXiv

  128. [138]

    The Scientist’s Rights to Research: A Constitutional Analysis

    John A Robertson. “The Scientist’s Rights to Research: A Constitutional Analysis”. In: S. Cal. l. Rev. 51 (1977), p. 1203

    1977

  129. [139]

    The Constitutional Status of American Science

    Steven Goldberg. “The Constitutional Status of American Science”. In: U. Ill. LF (1979), p. 1

    1979

  130. [140]

    Regulating scientific research: A constitutional mo- ment?

    Gert Verschraegen. “Regulating scientific research: A constitutional mo- ment?” In: Journal of Law and Society 45 (2018), S163–S184

    2018

  131. [141]

    Information hazards in biotechnology

    Gregory Lewis et al. “Information hazards in biotechnology”. In: Risk Analysis 39.5 (2019), pp. 975–981

    2019

  132. [142]

    SolidGoldMagikarp (plus, prompt generation)

    Jessica Rumbelow and Matt Watkins. SolidGoldMagikarp (plus, prompt generation). Tech. rep. LessWrong, February 5 2023

    2023

  133. [143]

    RAND Report: Securing AI Model Weights

    Sella Nevo et al. “RAND Report: Securing AI Model Weights”. In: (2024)

    2024

  134. [144]

    Conjecture Internal Infohazard Policy

    Connor Leahy et al. Conjecture Internal Infohazard Policy . Tech. rep. Conjecture, 29th July 2022

    2022

  135. [145]

    Man- aging Information Security Risk Organization, Mission, and Information System View

    Information Technology Laboratory Computer Security Division. Man- aging Information Security Risk Organization, Mission, and Information System View. Tech. rep. National Institute of Standards and Technology, 2017

    2017

  136. [146]

    GDPR Compliance in Cybersecurity Software: A Case Study of DPIA in Information Shar- ing Platform

    Martin Hor´ ak, V´ aclav Stupka, and Martin Hus´ ak. “GDPR Compliance in Cybersecurity Software: A Case Study of DPIA in Information Shar- ing Platform”. In: Proceedings of the 14th International Conference on Availability, Reliability and Security. ARES ’19. Canterbury, CA, U...

    2019

  137. [147]

    Understanding the (In)Effectiveness of Content Mod- eration: A Case Study of Facebook in the Context of the U.S

    Ian Goldstein et al. Understanding the (In)Effectiveness of Content Mod- eration: A Case Study of Facebook in the Context of the U.S. Capitol Riot

  138. [148]

    Experimental adaptation of an influenza H5 HA confers respiratory droplet transmission to a reassortant H5 HA/H1N1 virus in ferrets

    Masaki Imai et al. “Experimental adaptation of an influenza H5 HA confers respiratory droplet transmission to a reassortant H5 HA/H1N1 virus in ferrets”. In: Nature 486.7403 (2012), pp. 420–428

    2012

  139. [149]

    The 5 Most Important Revelations From the ‘Facebook Pa- pers’ https://time.com/6110234/facebook-papers-testimony-explained/

    Nik Popli. “The 5 Most Important Revelations From the ‘Facebook Pa- pers’ https://time.com/6110234/facebook-papers-testimony-explained/”. In: Time (October 26, 2021)

    2021

  140. [150]

    Glaze and the Effectiveness of Anti-AI Methods for Diffusion Models (https://huggingface.co/blog/parsee-mizuhashi/glaze-and- anti-ai-methods)

    Parsee Mizuhashi. Glaze and the Effectiveness of Anti-AI Methods for Diffusion Models (https://huggingface.co/blog/parsee-mizuhashi/glaze-and- anti-ai-methods). Tech. rep. HuggingFace, May 15 2024

    2024

  141. [151]

    Red teaming language models to reduce harms: Methods, scaling behaviors, and lessons learned

    Deep Ganguli et al. “Red teaming language models to reduce harms: Methods, scaling behaviors, and lessons learned”. In:arXiv preprint arXiv:2209.07858 (2022). 39

    2022 arXiv

  142. [152]

    Construction of an infectious horsepox virus vaccine from chemically synthesized DNA fragments

    Ryan S Noyce, Seth Lederman, and David H Evans. “Construction of an infectious horsepox virus vaccine from chemically synthesized DNA fragments”. In: PloS one 13.1 (2018), e0188453

    2018

  143. [153]

    Biodefense research: an emerging conundrum

    RM Atlas. “Biodefense research: an emerging conundrum”. In: Current Opinion Biotechnology (2005 June). 38

    2005

  144. [2023]

    arXiv: 2301.02737 [cs.SI]

    arXiv

  145. [2024]

    arXiv: 2401.04757 [cs.LG]

    arXiv

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.