Add quotes with AI Add quotes manually

Require open source AI models to include safety benchmarks in release notes

ai-ethics open-source policy research-policy ai-governance regulations ai-safety open-science ai-regulation transparency

Share this with your representatives, journalists and friends

Arguments for

Standardized safety documentation

Supporters say open-source releases should be accompanied by model cards, data cards, evaluation cards, or similar safety information. They argue that disclosing testing methodologies, evaluation criteria, results, limitations, and related safety details creates robust, industry-wide transparency and evaluation standards.

MCML is perfect for documenting open source models. Publish Model Cards and Dataset Cards alongside your model releases [...] Evaluation Card (benchmarking and testing)

MCML Standard · 2026 · full quote
[...] model cards support responsible AI development and the adoption of robust, industry-wide standards for broad transparency and evaluation practices

Google · 2025 · full quote
Disclose details such as testing methodologies, evaluation criteria, results, limitations, and gaps for any internal and external evaluations conducted prior to release.

Partnership on AI · Oct 24, 2023 · full quote
In November 2024, we authored a paper addressing the rapidly evolving ecosystem of publicly available AI models and their potential implications for security and safety. In this paper we proposed standardization of model cards and extensions to include safety, security, and data …

Huzaifa Sidhpurwala · Sep 3, 2025 · full quote
platforms that host libraries of open-source AI models to ensure that models are uploaded with appropriate system and safety information

Julia Smakman · 2025 · full quote

Openness enables independent testing

Proponents argue that open source makes models open to inspection, testing in varied settings, and even publicly verifiable development, so publishing benchmark results helps outsiders examine systems and contribute fixes. They emphasize independent testing and intelligible publication of results as a practical safety benefit of openness.

[T]he developer of an open source model could take a secure by design approach [...] train the model in a publicly verifiable way.

Cybersecurity and Infrastructure Security Agency (CISA) · 2024 · full quote
policymakers and regulators have no choice but to deal with both open and closed-source models. [...] If independent testing were done and results on an intelligible scale were published, I think that would encourage constructive competition.

Professor John McDermid · Nov 8, 2023 · full quote

No blanket exemption from safety

Supporters contend that open-source releases can still pose serious risks and therefore should not automatically be exempt from pre-deployment risk assessments, safety assurances, or public evaluation results. Some frame this most strongly for powerful or higher-risk open models, where methods, mitigations, and evaluations should be public before release.

Unsecured “open-source” AI systems pose a massive series of threats [...]. They deserve no exemptions and should be regulated just like other high-risk AI systems.

David Evan Harris · Dec 4, 2023 · full quote
Consider the potential risks of open-sourcing models. [...] publishing a transparency report (such as a model card) [...] risk assessment evaluation results.

Department for Science, Innovation and Technology. · Oct 27, 2023 · full quote
Congress should require standardized documentation, and like audited financial statements, they should be verifiable by a trusted third party, like an auditor, about the AI training data. [...] Congress should direct NIST to develop standards for high-stakes domains, such as heal…

Ramayya Krishnan · Sep 12, 2023 · full quote
For models with risk above the lowest level, most evaluation results and methods should be public, including any performed mitigations.

Jack Titus · Mar 5, 2024 · full quote
Policy and safeguards must evolve just as swiftly [...] requiring safety assurances for powerful open releases before the window for meaningful intervention closes.

Centre for Future Generations · 2026 · full quote

Arguments against

Regulate uses, not models

Opponents say policy should distinguish regulating applications from regulating the technology itself. From this perspective, benchmark mandates for open-source releases amount to regulating AI research, development, or the release of models, which they describe as unnecessary or ill-advised.

However, it is important to differentiate regulating applications (which we need) vs. regulating the technology (which is ill-advised).

Andrew Ng · Dec 11, 2023 · full quote
The question that people are debating is whether it makes sense to regulate research and development of AI. And I don't think it does.

Yann LeCun · 2023 · full quote
Regulating the open-source release of GPAI models is broadly unnecessary

Alex Engler · 2022 · full quote

Innovation and openness costs

Some critics argue that open source is central to software innovation and that they see benefits, not risks, in open-sourcing models. They warn that added obligations can slow open development and create significant, impractical barriers for open-source foundation models.

Open source is the heart of software innovation and this bill slows it down. [...] There should be zero liability at the model layer.

Anjney Midha · Jun 19, 2024 · full quote
So I don't see any risk associated with open sourcing models. I only see benefits.

Arthur Mensch · Aug 4, 2024 · full quote

Premature or unnecessary disclosures

Others say it is premature to require specific measurements and benchmarks before consensus exists, and some open-source standards views list model cards, evaluation code, validation datasets, and benchmarking datasets as likely not required. One legal formulation in the quotes also says obligations to provide evaluation results do not apply to providers using a free and open-source licence.

Until consensus develops, it is premature to require specific measurements and benchmarks.

Meta · Mar 26, 2024 · full quote
Providers [...] shall [...] the results of its evaluation. The obligations [...] shall not apply to providers of AI models [...] a free and open-source licence [...].

European Parliament and Council of the European Union · Jun 13, 2024 · full quote

Middle ground

Risk-based open-source duties

A narrower approach in the quotes would exempt many free and open-source releases, but impose extra evaluation and safety-benchmark duties on powerful, systemic-risk, or above-lowest-risk models. On this view, disclosure requirements turn on the model's risk level rather than applying uniformly to all open releases.

providers of open-source GPAISR models will be subject to additional safety and security obligations prescribed in Article 55(1a-d). [...] conduct evaluations using established safety benchmarks.

Yacine Jernite · full quote
For models with risk above the lowest level, most evaluation results and methods should be public, including any performed mitigations.

Jack Titus · Mar 5, 2024 · full quote
Policy and safeguards must evolve just as swiftly [...] requiring safety assurances for powerful open releases before the window for meaningful intervention closes.

Centre for Future Generations · 2026 · full quote

Insights

Supportive quotes frequently describe benchmark disclosure through documentation artifacts such as model cards, data cards, evaluation cards, transparency reports, and validation reports.
Several supporters treat openness as a safety asset because it enables inspection, testing, verification, and fixes, whereas several opponents treat openness as a reason for lighter regulation or exemptions.
Opposing quotes split between a principled objection to regulating the technology or model layer and a practical objection that fixed benchmark mandates are premature or burdensome.
Multiple quotes on both sides point toward risk differentiation, using terms such as powerful releases, models above the lowest risk level, and models with systemic risks.
The quoted voices span governments, companies, researchers, and standards-oriented organizations on both sides of the statement, showing the disagreement is not confined to one type of institution.

The quotes divide between those who see standardized documentation and published evaluations as a basic safety practice for open-source releases and those who view mandatory benchmark disclosure as overbroad regulation of the model layer. Several statements suggest a narrower risk-based approach for powerful or systemic-risk models, but there is no shared view in the quotes on whether most open-source models should be exempt or included.

Cast your vote:

Results (35 votes):

Total (35 votes)

For 26 (74%) Abstain 0 (0%) Against 9 (26%)

For (26)

Thinking Machines
AI model developer whose Inkling model card reports pre-release safety evaluations and benchmarks.

I am this person! Add a sourced quote Report comment

votes For and says:

Inkling reports reasoning, agentic, factuality, chat, vision, audio, and safety benchmark results at effort=0.99.
AI Verified source (Jul 14, 2026)

added 22d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Rebecca Finlay
CEO of Partnership on AI.

I am this person! Add a sourced quote Report comment

votes For and says:

Now we need to open that with more open benchmarks, verification, independent testing, making sure that models reflect linguistic diversity, open science, and technical standards.
AI Verified source (Jul 7, 2026)

added 23d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Independent International Scientific Panel on Artificial Intelligence
United Nations independent scientific panel that issued its July 2026 preliminary report on artificial intelligence.

I am this person! Add a sourced quote Report comment

votes For and says:

Researchers also consistently advocate for rigorous measurement and evaluation protocols to assess a model’s safety prior to publication. International bodies like the United Nations can play a critical role in coordinating global standards, ensuring...
more AI Verified source (Jul 2026)

added 25d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Hoki Kim
Professor at Chung-Ang University researching AI governance and open-weight foundation-model policy.

I am this person! Add a sourced quote Report comment

votes For and says:

The model card is restructured from prose README into a machine-readable structured document organized around three principles: (1) heritage transparency — full disclosure of training data, base model, and alignment lineage; (2) alignment provenance ...
more AI Verified source (May 7, 2026)

added 23d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Frontier Model Forum
Frontier AI industry forum

I am this person! Add a sourced quote Report comment

votes For and says:

Develop and publish monitorability evaluations. Standardized evaluations and metrics can improve assessments for how monitorable an AI model’s reasoning process is. These tests would measure the clarity, coherence, and faithfulness of a model’s Chain...
more AI Verified source (Jan 27, 2026)

added 1d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Huzaifa Sidhpurwala
Red Hat security engineer

I am this person! Add a sourced quote Report comment

votes For and says:

In November 2024, we authored a paper addressing the rapidly evolving ecosystem of publicly available AI models and their potential implications for security and safety. In this paper we proposed standardization of model cards and extensions to inclu...
more AI Verified source (Sep 3, 2025)

added 6mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Vincent Caldeira
CTO of Red Hat in APAC

I am this person! Add a sourced quote Report comment

votes For and says:

Class III (Open Model): The entry point for openness. This involves the partial release of components which typically includes the model architecture, final model weights, and basic documentation, like a technical report, evaluation results, a model ...
more AI Verified source (Jul 15, 2025)

added 6mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Julia Smakman
Legal researcher, Ada Lovelace Institute

I am this person! Add a sourced quote Report comment

votes For and says:

platforms that host libraries of open-source AI models to ensure that models are uploaded with appropriate system and safety information
AI Verified source (2025) 1 of 2

added 6mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Cheryl Seah
Director, Drew & Napier LLC

I am this person! Add a sourced quote Report comment

votes For and says:

Select models where [...] performance on evaluation and red-teaming efforts are disclosed, over models that don’t – model cards can help with providing this information;
AI Verified source (Nov 2024)

added 6mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Jack Titus
AI policy writer at FAS

I am this person! Add a sourced quote Report comment

votes For and says:

For models with risk above the lowest level, most evaluation results and methods should be public, including any performed mitigations.
AI Verified source (Mar 5, 2024)

added 6mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
David Evan Harris
UC Berkeley public scholar, tech policy advisor

I am this person! Add a sourced quote Report comment

votes For and says:

Unsecured “open-source” AI systems pose a massive series of threats [...]. They deserve no exemptions and should be regulated just like other high-risk AI systems.
AI Verified source (Dec 4, 2023)

added 6mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Professor John McDermid
Professor of safety-critical systems

I am this person! Add a sourced quote Report comment

votes For and says:

policymakers and regulators have no choice but to deal with both open and closed-source models. [...] If independent testing were done and results on an intelligible scale were published, I think that would encourage constructive competition.
AI Verified source (Nov 8, 2023)

added 6mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Irene Solaiman
Hugging Face chief policy officer, researcher

I am this person! Add a sourced quote Report comment

votes For and says:

For pre-deployment risk assessments for open source, [...] models, [...] are released with a model card. Model cards [...] will generally come with a set of evaluations.
AI Verified source (Nov 8, 2023)

added 6mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Department for Science, Innovation and Technology.
UK government department

I am this person! Add a sourced quote Report comment

votes For and says:

Consider the potential risks of open-sourcing models. [...] publishing a transparency report (such as a model card) [...] risk assessment evaluation results.
AI Verified source (Oct 27, 2023)

added 6mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Partnership on AI
AI governance nonprofit, multi-stakeholder

I am this person! Add a sourced quote Report comment

votes For and says:

Disclose details such as testing methodologies, evaluation criteria, results, limitations, and gaps for any internal and external evaluations conducted prior to release.
AI Verified source (Oct 24, 2023)

added 5mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Ramayya Krishnan
Dean, Heinz College, Carnegie Mellon

I am this person! Add a sourced quote Report comment

votes For and says:

Congress should require standardized documentation, and like audited financial statements, they should be verifiable by a trusted third party, like an auditor, about the AI training data. [...] Congress should direct NIST to develop standards for hig...
more AI Verified source (Sep 12, 2023)

added 6mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
HITRUST
AI assurance and certification organization

I am this person! Add a sourced quote Report comment

votes For and says:

The organization publishes model cards for the AI models it produces, which (minimally) include the following elements: [...] (4) evaluation details (e.g., methodology, metrics, outcomes).
AI Verified source

added 6mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Centre for Future Generations
European think tank on emerging technologies

I am this person! Add a sourced quote Report comment

votes For and says:

Policy and safeguards must evolve just as swiftly [...] requiring safety assurances for powerful open releases before the window for meaningful intervention closes.
Disputed source (2026)

added 5mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
MCML Standard
AI documentation standard initiative

I am this person! Add a sourced quote Report comment

votes For and says:

MCML is perfect for documenting open source models. Publish Model Cards and Dataset Cards alongside your model releases [...] Evaluation Card (benchmarking and testing)
Disputed source (2026)

added 5mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Julia Smakman
Legal researcher, Ada Lovelace Institute

I am this person! Add a sourced quote Report comment

votes For and says:

[...] platforms that host libraries of open-source AI models to ensure that models are uploaded with appropriate system and safety information
Disputed source (Dec 8, 2025)

added 6mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Google
Search and advertising company

I am this person! Add a sourced quote Report comment

votes For and says:

[...] model cards support responsible AI development and the adoption of robust, industry-wide standards for broad transparency and evaluation practices
Unverifiable source (2025)

added 5mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Cybersecurity and Infrastructure Security Agency (CISA)
US cybersecurity agency

I am this person! Add a sourced quote Report comment

votes For and says:

[T]he developer of an open source model could take a secure by design approach [...] train the model in a publicly verifiable way.
Disputed source (2024)

added 6mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
NVIDIA
GPU and AI computing company

I am this person! Add a sourced quote Report comment

votes For and says:

This highlighted the importance of disclosing key details regarding dataset demographic make-up, performance metrics for different demographic groups, and specific mitigation efforts [...].
Disputed source

added 6mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate

Abstain (0)

Against (7)

Arthur Mensch
Mistral AI co-founder and CEO

I am this person! Add a sourced quote Report comment

votes Against and says:

So I don't see any risk associated with open sourcing models. I only see benefits.
AI Verified source (Aug 4, 2024)

added 6mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Anjney Midha
a16z general partner

I am this person! Add a sourced quote Report comment

votes Against and says:

Open source is the heart of software innovation and this bill slows it down. [...] There should be zero liability at the model layer.
AI Verified source (Jun 19, 2024)

added 6mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Meta
Social media and AI company

I am this person! Add a sourced quote Report comment

votes Against and says:

Until consensus develops, it is premature to require specific measurements and benchmarks.
AI Verified source (Mar 26, 2024)

added 6mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Andrew Ng
Baidu; Stanford CS faculty; founded Coursera and Google Brain

I am this person! Add a sourced quote Report comment

votes Against and says:

However, it is important to differentiate regulating applications (which we need) vs. regulating the technology (which is ill-advised).
AI Verified source (Dec 11, 2023)

added 4mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Yann LeCun
Computer scientist, AI researcher

I am this person! Add a sourced quote Report comment

votes Against and says:

The question that people are debating is whether it makes sense to regulate research and development of AI. And I don't think it does.
AI Verified source (2023)

added 6mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Alex Engler
Brookings AI policy scholar

I am this person! Add a sourced quote Report comment

votes Against and says:

Regulating the open-source release of GPAI models is broadly unnecessary
Unverifiable source (2022)

added 10mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
European Parliament and Council of the European Union
EU co-legislators

I am this person! Add a sourced quote Report comment

votes Against and says:

Providers [...] shall [...] the results of its evaluation. The obligations [...] shall not apply to providers of AI models [...] a free and open-source licence [...].
Disputed source (Jun 13, 2024)

added 6mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate

Loading more opinions...

More h/ai-safety votes

If leading AI labs adopted a windfall clause (pre-committing to broadly share profits above a high threshold), it would meaningfully reduce competitive pressure to cut safety corners

Require large datacenters to install kill switches for AI containment

Pausing AI at human-genius level to solve alignment is safer than either racing to superintelligence or halting entirely

[🇺🇸 Congress, Oct 2025] Requiring age verification and content safeguards for AI chatbots to protect minors

The U.S. and China should establish Mutually Assured Compute Destruction

Back