mlsanityData to 14 Oct 2026

The Gradient of Generative AI Release: Methods and Considerations

arXiv 2302.04844 ↗First posted 5 Feb 2023cs.CY cs.AIcited by 63

Not on the map. Why?

Abstract (arXiv)

As increasingly powerful generative AI systems are developed, the release method greatly varies. We propose a framework to assess six levels of access to generative AI systems: fully closed; gradual or staged access; hosted access; cloud-based or API access; downloadable access; and fully open. Each level, from fully closed to fully open, can be viewed as an option along a gradient. We outline key considerations across this gradient: release methods come with tradeoffs, especially around the tension between concentrating power and mitigating risks. Diverse and multidisciplinary perspectives are needed to examine and mitigate risk in generative AI systems from conception to deployment. We show trends in generative system release over time, noting closedness among large companies for powerful systems and openness among organizations founded on principles of openness. We also enumerate safety controls and guardrails for generative systems and necessary investments to improve future releases.

Key facts. Cited by 63 papers in the map (data through 14 Oct 2026). Builds on 47 of its 116 references that are in the map. First posted on arXiv 5 Feb 2023.

Builds on

47 of 116 references.

References · 47

  1. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,863
  2. 2022Training language models to follow instructions with human feedbackcited by 15,971
  3. 2022Hierarchical Text-Conditional Image Generation with CLIP Latentscited by 6,585
  4. 2022Photorealistic Text-to-Image Diffusion Models with Deep Language Understandingcited by 6,355
  5. 2022PaLM: Scaling Language Modeling with Pathwayscited by 5,326
  6. 2021On the Opportunities and Risks of Foundation Modelscited by 4,033
  7. 2022OPT: Open Pre-trained Transformer Language Modelscited by 3,512
  8. 2020The Pile: An 800GB Dataset of Diverse Text for Language Modelingcited by 2,411
  9. 2020Extracting Training Data from Large Language Modelscited by 2,189
  10. 2022Make-A-Video: Text-to-Video Generation without Text-Video Datacited by 1,742
  11. 2018Datasheets for Datasetscited by 1,386
  12. 2022LaMDA: Language Models for Dialog Applicationscited by 1,220
35 more
  1. 2021Scaling Language Models: Methods, Analysis & Insights from Training Gophercited by 1,202
  2. 2018Model Cards for Model Reportingcited by 1,080
  3. 2021Ethical and social risks of harm from Language Modelscited by 966
  4. 2020Language (Technology) is Power: A Critical Survey of "Bias" in NLPcited by 956
  5. 2016Semantics derived automatically from language corpora contain human-like biasescited by 901
  6. 2023A Watermark for Large Language Modelscited by 725
  7. 2022GPT-NeoX-20B: An Open-Source Autoregressive Language Modelcited by 713
  8. 2020CrowS-Pairs: A Challenge Dataset for Measuring Social Biases in Masked Language Modelscited by 710
  9. 2019Release Strategies and the Social Impacts of Language Modelscited by 513
  10. 2019GLTR: Statistical Detection and Visualization of Generated Textcited by 496
  11. 2020Closing the AI Accountability Gap: Defining an End-to-End Framework for Internal Algorithmic Auditingcited by 384
  12. 2021AI and the Everything in the Whole Wide World Benchmarkcited by 349
  13. 2022Red-Teaming the Stable Diffusion Safety Filtercited by 264
  14. 2021Process for Adapting Language Models to Society (PALMS) with Values-Targeted Datasetscited by 181
  15. 2019Gradio: Hassle-Free Sharing and Testing of ML Models in the Wildcited by 178
  16. 2021The Values Encoded in Machine Learning Researchcited by 172
  17. 2020Automatic Detection of Machine Generated Text: A Critical Surveycited by 166
  18. 2018The Malicious Use of Artificial Intelligence: Forecasting, Prevention, and Mitigationcited by 165
  19. 2020Toward Trustworthy AI Development: Mechanisms for Supporting Verifiable Claimscited by 165
  20. 2023Generative Language Models and Automated Influence Operations: Emerging Threats and Potential Mitigationscited by 158
  21. 2019Robust Invisible Video Watermarking with Attentioncited by 149
  22. 2020Decolonial AI: Decolonial Theory as Sociotechnical Foresight in Artificial Intelligencecited by 101
  23. 2020The Radicalization Risks of GPT-3 and Advanced Neural Language Modelscited by 76
  24. 2022Outsider Oversight: Designing a Third Party Audit Ecosystem for AI Governancecited by 67
  25. 2022Interactive Model Cards: A Human-Centered Approach to Model Documentationcited by 61
  26. 2020A Systematic Review on Model Watermarking for Neural Networkscited by 52
  27. 2020Behavioral Use Licensing for Responsible AIcited by 46
  28. 2022Structured access: an emerging paradigm for safe AI deploymentcited by 41
  29. 2022What Do NLP Researchers Believe? Results of the NLP Community Metasurveycited by 28
  30. 2022A Hazard Analysis Framework for Code Synthesis Large Language Modelscited by 22
  31. 2020Limits of Detecting Text Generated by Large-Scale Language Modelscited by 15
  32. 2019Reducing malicious use of synthetic media research: Considerations and potential release practices for machine learningcited by 12
  33. 2019The Offense-Defense Balance of Scientific Knowledge: Does Publishing AI Research Reduce Misuse?cited by 12
  34. 2019The tension between openness and prudence in AI researchcited by 5
  35. 2019Let's Keep It Safe: Designing User Interfaces that Allow Everyone to Contribute to AI Safetycited by 1

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2023: 22; 2024: 22; 2025: 13; 2026: 6.

63 papers.

Sort:

By area:

  1. 2023StarCoder: may the source be with you!cited by 799
  2. 2024Black-Box Access is Insufficient for Rigorous AI Auditscited by 126
  3. 2023Frontier AI Regulation: Managing Emerging Risks to Public Safetycited by 98
  4. 2023Evaluating the Social Impact of Generative AI Systems in Systems and Societycited by 92
  5. 2024Red-Teaming for Generative AI: Silver Bullet or Security Theater?cited by 83
  6. 2024Visibility into AI Agentscited by 79
  7. 2024On the Societal Impact of Open Foundation Modelscited by 65
  8. 2023Opening up ChatGPT: Tracking openness, transparency, and accountability in instruction-tuned text generatorscited by 56
  9. 2023Open-Sourcing Highly Capable Foundation Models: An evaluation of risks, benefits, and alternative methods for pursuing open-source objectivescited by 31
  10. 2023Democratising AI: Multiple Meanings, Goals, and Methodscited by 29
  11. 2023AI capabilities can be significantly improved without expensive retrainingcited by 28
  12. 2024The Model Openness Framework: Promoting Completeness and Openness for Reproducibility, Transparency, and Usability in Artificial Intelligencecited by 23
38 more
  1. 2023What does it take to catch a Chinchilla? Verifying Rules on Large-Scale Neural Network Training via Compute Monitoringcited by 21
  2. 2023Language as Reality: A Co-Creative Storytelling Game Experience in 1001 Nights using Generative AIcited by 21
  3. 2023Position: Key Claims in LLM Research Have a Long Tail of Footnotescited by 20
  4. 2023Bias and Fairness in Chatbots: An Overviewcited by 20
  5. 2024The 2024 Foundation Model Transparency Indexcited by 19
  6. 2024The AI Community Building the Future? A Quantitative Analysis of Development Activity on Hugging Face Hubcited by 18
  7. 2023A Cost Analysis of Generative Language Models and Influence Operationscited by 17
  8. 2023Protecting Society from AI Misuse: When are Restrictions on Capabilities Warranted?cited by 16
  9. 2023Safety and Fairness for Content Moderation in Generative Modelscited by 16
  10. 2024Harm Amplification in Text-to-Image Modelscited by 15
  11. 2025Deepfakes on Demand: the rise of accessible non-consensual deepfake image generatorscited by 15
  12. 2024Near to Mid-term Risks and Opportunities of Open-Source Generative AIcited by 13
  13. 2024Risks and Opportunities of Open-Source Generative AIcited by 13
  14. 2024Acceptable Use Policies for Foundation Modelscited by 12
  15. 2025The 2025 Foundation Model Transparency Indexcited by 11
  16. 2024Societal Adaptation to Advanced AIcited by 10
  17. 2023GPT-SW3: An Autoregressive Language Model for the Nordic Languagescited by 8
  18. 2025Attestable Audits: Verifiable AI Safety Benchmarks Using Trusted Execution Environmentscited by 8
  19. 2023Hazards from Increasingly Accessible Fine-Tuning of Downloadable Foundation Modelscited by 7
  20. 2025MusGO: A Community-Driven Framework For Assessing Openness in Music-Generative AIcited by 7
  21. 2023Exploring the Relevance of Data Privacy-Enhancing Technologies for AI Governance Use Casescited by 6
  22. 2024Sociotechnical Implications of Generative Artificial Intelligence for Information Accesscited by 6
  23. 2024From Silos to Systems: Process-Oriented Hazard Analysis for AI Systemscited by 6
  24. 2023Towards Responsible Governance of Biological Design Toolscited by 5
  25. 2024How far can bias go? Tracing bias from pretraining data to alignmentcited by 5
  26. 2023To Build Our Future, We Must Know Our Past: Contextualizing Paradigm Shifts in Natural Language Processingcited by 4
  27. 2023Transductive Learning for Textual Few-Shot Classification in API-based Embedding Modelscited by 4
  28. 2025Misalignment or misuse? The AGI alignment tradeoffcited by 4
  29. 2023Thorny Roses: Investigating the Dual Use Dilemma in Natural Language Processingcited by 3
  30. 2024The Implications of Open Generative Models in Human-Centered Data Science Work: A Case Study with Fact-Checking Organizationscited by 3
  31. 2025Modeling the Economic Impacts of AI Openness Regulationcited by 3
  32. 2024Governance of Generative Artificial Intelligence for Companiescited by 2
  33. 2024Leveraging Open-Source Large Language Models for Native Language Identificationcited by 2
  34. 2024Human Bias in the Face of AI: Examining Human Judgment Against Text Labeled as AI Generatedcited by 2
  35. 2025Audits Under Resource, Data, and Access Constraints: Scaling Laws For Less Discriminatory Alternativescited by 2
  36. 2026How Open Must Language Models be to Enable Reliable Scientific Inference?cited by 2
  37. 2023Contextual Confidence and Generative AIcited by 1
  38. 2024GPT, Ontology, and CAABAC: A Tripartite Personalized Access Control Model Anchored by Compliance, Context and Attributecited by 1

50 most cited of 63 All 63 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2023Towards Understanding the Interplay of Generative Artificial Intelligence and the Internet
  2. 2024A Formal Framework for Assessing and Mitigating Emergent Security Risks in Generative AI Models: Bridging Theory and Dynamic Risk Mitigation
  3. 2025SAIF: A Comprehensive Framework for Evaluating the Risks of Generative AI in the Public Sector
  4. 2023Exploring the intersection of Generative AI and Software Development
  5. 2023Identifying and Mitigating the Security Risks of Generative AI
  6. 2026Beyond the Binary: A nuanced path for open-weight advanced AI ↗
  7. 2026From AI-Generated Content to Agentic Action: Security and Safety Threats in Generative AI
  8. 2023On the Trustworthiness Landscape of State-of-the-art Generative Models: A Survey and Outlook