open-source ai-risk ai-governance policy ai-safety ai-regulation ethics

Share this with friends, journalists and politicians

Open-source AI is more dangerous than closed-source AI

Key insights

Much of the disagreement centers specifically on downloadable model weights rather than on open standards or source code alone.
Both sides treat advanced AI as dual-use technology, but they disagree on whether wider access chiefly empowers malicious actors or also equips defenders.
Supporters of the statement emphasize the loss of provider control after release, while opponents emphasize the security value of outside inspection, auditing, and shared remediation.
Capability level and timing repeatedly shape the debate: several quotes focus on open models that are only months behind the frontier, while others say lower-capability open models can still be beneficial.
Several quotes reject a simple all-or-nothing framing and instead call for testing, targeted safeguards, or model-specific release choices.

Arguments for

Irreversible loss of control

Once model weights are public, developers cannot recall them, patch existing copies, monitor use, or intervene after misuse. Proponents say safety layers can be stripped with little effort, leaving persistent harmful variants in circulation.

For models with dangerous capabilities – including highly cyber-capable models – open weight release therefore creates a persistent and irreversible risk of misuse.

AI Security Institute (AISI) · Jul 17, 2026 · full quote
You can’t patch a released, open-weight model. This is the property that breaks the assumptions of the prior model of responsible disclosure. Open-source software can be patched in place; a deployer updates and the hazard closes. An open-weight model cannot. A new version is a ne…

MLCommons · Jun 9, 2026 · full quote
Open weights make that worry land harder, because you can't recall them. A closed model that misbehaves can be patched or pulled overnight. A set of weights on a hard drive is forever. And the safety training is the flimsiest layer in the stack. You can fine-tune the refusals off…

Om Bharatiya · Jul 10, 2026 · full quote
Because open-weight safeguards can often be removed, public release of such models may make it impossible to prevent a wide range of actors from accessing dangerous capabilities.

Spencer E. Amdur · Jul 11, 2026 · full quote
When a frontier lab releases model weights publicly, the decision is irreversible in a way that no other software release is: the weights can be copied, fine-tuned, stripped of safety mitigations, and redistributed before any downstream harm is even detected.

Marisol Vega · Feb 19, 2026 · full quote
However, compared to closed-weight models, open-weight models are more easily modified to exhibit potentially harmful behaviours, and monitoring their usage is more difficult.

International AI Safety Report 2026 · Feb 2026 · full quote

Low barriers to malicious adaptation

Open models can be run locally without oversight, fine-tuned for specific harmful tasks, and cheap attacks can remove safeguards. Proponents argue this enables lone actors, criminals, and other harmful users to adapt models rapidly and deploy them at scale.

It does. And I’m a huge believer in open source and all the rest. But the issue is it’s a dangerous piece of technology that, if you can build it in an isolated environment and run in an isolated environment [...] one to three people in their garage [...] could take an open model…

Reid Hoffman · Sep 23, 2026 · full quote
In this paper, we show that open-weight safeguards are susceptible to simpler strategies that, despite being well known, have not been systematically evaluated against these safeguards. Specifically, we evaluate two low-cost attacks--abliteration and prefilling--that do not rely …

Kevin Kuo · May 26, 2026 · full quote
An attacker can run it locally without safety guardrails, fine-tune it against their specific targets, and operate with zero visibility to any provider or defender.

Travis Lanham · Jun 25, 2026 · full quote
Stripping guardrails from an open-source AI model is trivial for anyone who wants both.

Rob T. Lee · Jul 28, 2026 · full quote
Open models have a tendency to be less safe a lot of the time than closed ones. Maybe it could be misused by someone who downloads it and modifies it to remove safeguards or to sharpen its harmful capabilities. This type of fine-tuning could cost hundreds or a few thousand dollar…

Stephen Casper · Oct 19, 2026 · full quote
The most interesting risk in AI is not the OpenAI and Anthropic models. It's open-source models. They're way more dangerous because they can't be monitored.

Vinod Khosla · Oct 1, 2026 · full quote

Near-frontier capabilities in public hands

Several quotes argue that open-weight systems are only months behind leading closed models on cyber and biological capabilities, and are sometimes released with weaker evaluations or safeguards. As capability rises, they say, broadly available access puts increasingly powerful tools into many hands.

On biological capabilities, independent evaluator SecureBio found that Kimi K3 was roughly comparable to closed models from about eight months earlier. When prompted with questions designed to elicit dangerous biological information, the model declined to answer only 26.9% of the…

Kar Mun Nicole Wong · Oct 14, 2026 · full quote
Open-weight models are approaching the capabilities of frontier closed-weight models, yet they are often released without the safety evaluations that accompany closed models, and their safeguards are either weak or can be easily removed.

SaferAI · Aug 2, 2026 · full quote
Open-weight models – models of which the core parameters are made available for others to run and adapt – are an important medium-term issue. Evidence suggests they may lag the most advanced closed models by only four to eight months, and that their safeguards are often easy to r…

Bank of England · Jul 1, 2026 · full quote
Open-weight releases such as Kimi K2.5 are already benchmarked on par with frontier closed models on cybersecurity tasks. Safeguards on open-weight models can also be removed through fine-tuning, a well-established procedure.

CERT-EU · Jul 2026 · full quote
According to common benchmarks, open weight models are estimated to be 3 to 12 months behind the frontier (varying across capabilities). Developers of frontier closed models have recently designated these closed models as ‘High’ risk for cyber and biological misuse, and added saf…

The 2026 Singapore Consensus on Global AI Safety Research Priorities · Jul 2026 · full quote
Because they can also be run more cheaply than proprietary models, the point at which open-weight models become dangerous and practical to use may arrive before they fully catch up with frontier models.

Financial Policy Committee · Jul 15, 2026 · full quote

Closed access enables intervention

Supporters of the statement argue that closed models can be gated, monitored, patched, or withdrawn, while open-weight copies cannot. They present withholding open release as a concrete defense against downstream cyber, biological, and criminal misuse.

Open-weight models introduce new and unique risks – they provide near-frontier capabilities which could be used for criminal purposes, are widely available to anyone who intends harm, and are easier to misuse. [...] Major frontier AI companies such as OpenAI, Anthropic and Google…

George Balston · Sep 18, 2026 · full quote
Not deploying high-risk systems with open weights: Avoiding open-weight release of a system can be a strong defence against many downstream risks that they pose.

UK AI Security Institute · Jul 22, 2026 · full quote
Higher security: Closed models have more safeguards. However, they also make certain safety measures harder, such as red teaming and mechanistic interpretability. Lower security: Open-weight systems are harder to monitor and regulate and can easily be fine-tuned to remove safegua…

Government of the United Kingdom · Jun 2026 · full quote
For closed-weight models, [misuse] risk is at least partially manageable. Input-output classifiers and safety fine-tuning can intercept most non-sophisticated actors before they get actionable guidance. [...] Open-weight models have no such layer. Once weights are downloaded, the…

Sophie Kim · Apr 11, 2026 · full quote
Compounding the problem, labs in China often release dual-use capable models as open-weight. Once a model is open-weight, safeguards that do exist can be removed, making the model available to any state or non-state actor to use for malicious purposes, including the cyber and CBR…

Anthropic · May 14, 2026 · full quote
The same models that hunt for software flaws can, without the right safeguards, help a bad actor work through the hardest steps of building a biological weapon. Our leading laboratories build in guardrails to refuse that kind of help, but when a foreign adversary copies an Americ…

Andy Ogles · Jun 4, 2026 · full quote

Arguments against

Transparency strengthens security

Opponents argue that openness allows independent inspection, auditing, red teaming, and faster identification of flaws or unwanted behaviors. They say black-box systems ask the public to trust a provider instead of letting a broader community test and improve the technology.

As policymakers and regulators work through questions of AI safety, the distinction that matters is not open versus closed, but transparent and secure versus opaque and unexamined. Open models and open tooling should be recognized as defensive assets that enable transparency, ind…

The Linux Foundation · Jul 27, 2026 · full quote
Safe open-weight models are public goods, as they put AI development and safety work in many hands and make training choices inspectable.

Thinking Machines Lab · Jul 31, 2026 · full quote
Open source does not eliminate risk. It changes how risk is managed. It allows more researchers, developers, and defenders to examine systems, test assumptions, surface weaknesses, and harden code under real-world conditions. [...] If frontier models are increasingly capable of f…

Rob Thomas · 2026 · full quote
Openness also delivers genuine transparency, because when the model is yours to inspect, you can see exactly what sits inside it, [...] and turn your best and brightest loose to hunt the very security flaws the cyber-hawks rightly fear. A sealed model offers no such window and as…

Jordan Schachtel · Jul 6, 2026 · full quote
Open-source AI allows the public to audit and verify algorithms, [...] It is also more secure because any attacks or exploits can be identified and remediated.

Travis Oliphant · Feb 21, 2025 · full quote
Transparency is the greatest safeguard. [...] When codes are open and shared, millions of young minds can make them safer and better.

Narendra Modi · Feb 19, 2026 · full quote

Openness is not the main variable

A common counterargument is that risk turns on capability, use case, deployment, and safeguards rather than on whether weights are open or closed. On this view, advanced AI can be dangerous in either form, and safety policy should be risk-based instead of treating openness as inherently worse.

Weight openness alone does not determine an AI system’s risk. Risk depends on the model, use case, deployment, and safeguards. Open- and closed-weight models can share many of the same underlying risks, but how those risks materialize depends on the deployment.

Michael Scovetta · Oct 16, 2026 · full quote
Open models, like any powerful technology, can be misused — including through attempts to weaken safeguards or repurpose capabilities for cyber attacks — but those risks are not unique to open systems, and they must be managed wherever advanced AI is deployed.

NVIDIA · Jul 27, 2026 · full quote
Open-weight systems are not inherently less secure than closed systems.

John Villasenor · Aug 10, 2026 · full quote
If safeguards are properly followed, businesses, including banks, should have enough guarantees to essentially run any model, regardless of where they come from.

Marco Argenti · Sep 8, 2026 · full quote
From a business perspective, the difference between open and closed is a little bit overblown. From a security perspective, the difference between open and closed models is, for some intents and purposes, overblown. The most important thing is how powerful a model is. If a model …

Dario Amodei · 2025 · full quote
Study of 34 AI models from 10 global providers finds open-source AI models are not less safe, reasoning models are hardest to exploit and smaller models are the most vulnerable.

TELUS Digital · May 26, 2026 · full quote

Defensive access matters

Opponents say defenders need broad, timely access to strong models because the same capabilities that can aid attackers can also support incident response, enterprise security, and public-sector protection. Restricting open models, they argue, can leave smaller defenders dependent on closed providers or without equivalent tools.

Open-weight models have been a boon for defenders. When Hugging Face investigated how OpenAI-powered agents penetrated its systems, proprietary models refused some of the forensic work, but the Chinese open-weight model GLM-5.2 helped reconstruct the intrusion. [...] Open-weight …

The Washington Post Editorial Board · Oct 25, 2026 · full quote
When a frontier model is attacking you and moving laterally inside your infrastructure, defenders need wide access to near-frontier tools within hours or even minutes, rather than being pointed toward a closed-door platform.

Thomas Wolf · Jul 22, 2026 · full quote
Shared access to open-source AI systems is important for protection in a world where AI can be used to hack into electronic systems, like voting systems and benefits infrastructure.

Matt Lane · Aug 10, 2026 · full quote
Open models, a valuable defense tool, are today being tuned, evaluated, and run in production to actively improve the security of enterprises.

Bartley Richardson · Jul 27, 2026 · full quote
Open-source AI is more than a catalyst for commercial innovation. It is a strategic advantage for the U.S. military, enabling American companies to push the technology forward while giving the DOD a foundation it can inspect, secure, adapt and deploy faster than any adversary.

Patrick McSpadden · Jul 16, 2026 · full quote
The alliance used its launch to urge regulators to treat open models as ‘defensive assets, not liabilities,’ warning that blanket restrictions would weaken cyber defenses and concentrate power in a handful of closed AI providers.

Open Secure AI Alliance · Jul 28, 2026 · full quote

Closed concentration can be dangerous

Several opponents argue that concentrating advanced AI inside a few proprietary systems creates its own security and governance risks, including excessive power over information and reduced checks and balances. They portray broader access and competition as a safeguard against dependence on a small number of companies.

One of the ways to achieve both individual empowerment and checks and balances is to make sure you put the technology in everybody's hands. I’m personally more worried about centralization than I am about any of the specific risks others are talking about.

Mark Zuckerberg · Aug 10, 2026 · full quote
I see the danger of this concentration of power through proprietary AI systems as a much bigger danger than everything else. What works against this is people who think that for reasons of security, we should keep AI systems under lock and key because it’s too dangerous to put it…

Yann LeCun · 2024 · full quote
But the real danger lies in centralized, black-box systems that resist open auditing.

Himanshu Tyagi · 2025 · full quote
With the right choices, open weight AI can expand opportunity, strengthen competition, extend American technological leadership, mitigate risk, and ensure that the benefits of this extraordinary technology are shared broadly across our economy.

Open Weights and American AI Leadership · Jul 24, 2026 · full quote
And that's where I think that that first debate around open-source AI being riskier than closed AI got it wrong and frankly was instrumentalized by a certain set of players.

Mark Surman · Apr 1, 2026 · full quote

Middle ground

Capability thresholds change the calculus

Some participants say lower-capability open models can be a net benefit, but that the case changes as systems become powerful enough for serious misuse. They support testing and different release bars based on capability, with some dangerous systems deemed unsuitable for open release even if closed access remains possible.

Below middling levels of capability, probably a net benefit. As they become more capable, open-weight models become much more dangerous than closed-weight models. For instance, imagine if five billion people had free access to an open-weight model more powerful than Claude Mythos…

Stuart J. Russell · Sep 14, 2026 · full quote
The right way to do this is if a model, before it is released in open weights or in other ways, a model should be tested. Can it be used by people? Are we sure we’re not going to lose control of it? Does it have the capabilities or intentions that could be dangerous for people? I…

Yoshua Bengio · Mar 12, 2026 · full quote
From a business perspective, the difference between open and closed is a little bit overblown. From a security perspective, the difference between open and closed models is, for some intents and purposes, overblown. The most important thing is how powerful a model is. If a model …

Dario Amodei · 2025 · full quote
Not deploying high-risk systems with open weights: Avoiding open-weight release of a system can be a strong defence against many downstream risks that they pose.

UK AI Security Institute · Jul 22, 2026 · full quote
Muse Spark had triggered safety checks that made it "not suitable for open sourcing."

Alexandr Wang · May 13, 2026 · full quote

Targeted release design

Another nuanced view is that the safety impact of restricting access depends on whether it slows attackers more than defenders and on what governed alternatives exist. These quotes point to options such as controlled access, defender-first windows, focused safeguards, or concern focused specifically on stripped-down variants rather than all open models.

Restricting access to a dual-use AI model is precautionary only if it delays harmful actors more than defenders. That condition varies across actors: a state agency or organized criminal group may obtain a substitute through theft, distillation, intermediated access, independent …

Daniel Commey · Jul 24, 2026 · full quote
Open weights are not the problem in themselves. The acute and least reversible risk is the abliterated build, an open model with its safety training stripped out: the two we tested complied with 89% and 100% of requests, and once downloaded they cannot be recalled.

Adam Hadley · Jul 2026 · full quote
The governance of open-weight artificial intelligence (AI) models has been framed as a binary choice: openness as risk, restriction as safety. This paper challenges that framing, arguing that access restrictions, without governed alternatives, may displace risks rather than reduc…

Vinicius Santana Gomes · Apr 19, 2026 · full quote
Like policymakers, we want a safe, secure AI ecosystem, and view openness as an important part of achieving it. We support policies that preserve the ability to develop, distribute, and use open weight models subject to focused, risk-based safeguards and well-targeted tools for a…

GitLab · Jul 29, 2026 · full quote
It is unquestionably both. We can use the frontier models but also open weight models to prove a solution which gives defenders powerful capabilities and on their terms.

Jaya Baloo · Jun 22, 2026 · full quote

This debate is less about openness in the abstract than about what happens when increasingly capable model weights can be downloaded, modified, and run outside provider control. Participants sharply disagree on whether that prospect chiefly enlarges irreversible misuse or whether restricting it would weaken transparency, defensive capacity, and checks on concentrated power. Several quotes also point toward model-specific, threshold-based release policies rather than a single rule for all open systems.

Cast Your Vote:

Results (121 votes):

Total (121 votes)

For 61 (50%) Abstain 3 (2%) Against 57 (47%)

New Top

For (57)

Stephen Casper
AI safety researcher at the UK AI Security Institute, researching safeguards for open-weight AI models.

I am this person! Add a sourced quote Report comment

votes For and says:

Open models have a tendency to be less safe a lot of the time than closed ones. Maybe it could be misused by someone who downloads it and modifies it to remove safeguards or to sharpen its harmful capabilities. This type of fine-tuning could cost hun...
more AI Verified source (Oct 19, 2026)

added 2d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Maxime Stauffer
Co-founder and CEO of the Simon Institute for Longterm Governance, a Geneva think tank focused on international governance of frontier AI.

I am this person! Add a sourced quote Report comment

votes For and says:

If you have an extremely powerful capability, but no one has access to it and it remains under control, then the risk is almost zero. But if everyone has access to it, then the risk is likely higher.
AI Verified source (Oct 15, 2026)

added 9d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Kar Mun Nicole Wong
AI governance researcher; Yale School of Management and Jackson School of Global Affairs alumna.

I am this person! Add a sourced quote Report comment

votes For and says:

On biological capabilities, independent evaluator SecureBio found that Kimi K3 was roughly comparable to closed models from about eight months earlier. When prompted with questions designed to elicit dangerous biological information, the model declin...
more AI Verified source (Oct 14, 2026)

added 13d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Guillaume Princen
Anthropic's head of international.

I am this person! Add a sourced quote Report comment

votes For and says:

We need to be really careful with open-source on safety. I can't talk for open models. But by definition, there's more risk with them.
AI Verified source (Oct 14, 2026)

added 17d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Vinod Khosla
Venture capitalist; Khosla Ventures founder

I am this person! Add a sourced quote Report comment

votes For and says:

The most interesting risk in AI is not the OpenAI and Anthropic models. It's open-source models. They're way more dangerous because they can't be monitored.
AI Verified source (Oct 1, 2026)

added 28d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Reid Hoffman
LinkedIn co-founder, venture capitalist, author of Superagency

I am this person! Add a sourced quote Report comment

votes For and says:

It does. And I’m a huge believer in open source and all the rest. But the issue is it’s a dangerous piece of technology that, if you can build it in an isolated environment and run in an isolated environment [...] one to three people in their garage ...
more AI Verified source (Sep 23, 2026)

added 2d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
George Balston
Strategic Adviser to the Alan Turing Institute and co-founder of AVERI; former UK government researcher, data scientist, and technical lead focused on AI, privacy, and government technology.

I am this person! Add a sourced quote Report comment

votes For and says:

Open-weight models introduce new and unique risks – they provide near-frontier capabilities which could be used for criminal purposes, are widely available to anyone who intends harm, and are easier to misuse. [...] Major frontier AI companies such a...
more AI Verified source (Sep 18, 2026)

added 1mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Stuart J. Russell
AI Expert and Professor

I am this person! Add a sourced quote Report comment

votes For and says:

Below middling levels of capability, probably a net benefit. As they become more capable, open-weight models become much more dangerous than closed-weight models. For instance, imagine if five billion people had free access to an open-weight model mo...
more AI Verified source (Sep 14, 2026)

added 7d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Sirisha Rambhatla
Professor of management science and engineering at the University of Waterloo; director of its Critical Machine Learning Lab.

I am this person! Add a sourced quote Report comment

votes For and says:

When the safety guardrails are stripped out of a capable model, it can be used at scale for harm in ways a single person could never manage manually. [...] The leading open-weight models are often not too far behind the best closed models. As they gr...
more AI Verified source (Aug 25, 2026)

added 1mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Z.ai
Chinese artificial intelligence company behind the GLM family of language models.

I am this person! Add a sourced quote Report comment

votes For and says:

No safety system can eliminate every dual-use risk. Once model weights are public, no developer can guarantee control over every downstream modification or use. Model-level safeguards can raise the barrier to abuse, but they cannot provide absolute c...
more AI Verified source (Aug 16, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Danny Jenkins

I am this person! Add a sourced quote Report comment

votes For and says:

That assumption is unrealistic in the world of AI-driven automation and open-source agent ecosystems.
AI Verified source (Aug 5, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Rebecca Bellan
Senior reporter at TechCrunch covering AI business, policy, and emerging trends.

I am this person! Add a sourced quote Report comment

votes For and says:

But the safeguards in place for closed models don’t work at all on open-weight models, which are designed to run on any infrastructure with any set of safeguards — or lack thereof.
AI Verified source (Aug 4, 2026)

added 1mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Le Monde
French daily newspaper; editorial board

I am this person! Add a sourced quote Report comment

votes For and says:

So-called "open-weight" models, whose parameters can be downloaded and modified, were thought to be more likely to escape the protections put in place by their creators.
AI Verified source (Aug 4, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
SaferAI
Nonprofit organization researching and promoting responsible frontier AI risk management and safety governance.

I am this person! Add a sourced quote Report comment

votes For and says:

Open-weight models are approaching the capabilities of frontier closed-weight models, yet they are often released without the safety evaluations that accompany closed models, and their safeguards are either weak or can be easily removed.
AI Verified source (Aug 2, 2026)

added 1mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Frank Teruel
Chief Operating Officer at Arkose Labs, commenting on AI-enabled cyber threats.

I am this person! Add a sourced quote Report comment

votes For and says:

Frontier labs can put classifiers in front of these capabilities, but open-weight models a few months behind cannot, and the copies already downloaded never will.
AI Verified source (Jul 28, 2026)

added 3mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Alexander Leslie
Senior adviser at Recorded Future, commenting on AI-enabled cyber threats.

I am this person! Add a sourced quote Report comment

votes For and says:

Don't be surprised if these capabilities are soon in the hands of ransomware crews, intelligence agencies and a "lone operator renting compute by the hour."
AI Verified source (Jul 28, 2026)

added 3mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate

Loading more opinions...

Abstain (3)

Loading more opinions...

Against (56)

The Washington Post Editorial Board
Editorial board of The Washington Post, publishing the newspaper’s institutional opinions.

I am this person! Add a sourced quote Report comment

votes Against and says:

Open-weight models have been a boon for defenders. When Hugging Face investigated how OpenAI-powered agents penetrated its systems, proprietary models refused some of the forensic work, but the Chinese open-weight model GLM-5.2 helped reconstruct the...
more AI Verified source (Oct 25, 2026)

added 8d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Michael Scovetta
Partner Security Assurance Manager at Microsoft, specializing in AI and open-source security.

I am this person! Add a sourced quote Report comment

votes Against and says:

Weight openness alone does not determine an AI system’s risk. Risk depends on the model, use case, deployment, and safeguards. Open- and closed-weight models can share many of the same underlying risks, but how those risks materialize depends on the ...
more AI Verified source (Oct 16, 2026)

added 7d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Marco Argenti
Chief Information Officer of Goldman Sachs, speaking publicly on safeguards for open-weight AI models.

I am this person! Add a sourced quote Report comment

votes Against and says:

If safeguards are properly followed, businesses, including banks, should have enough guarantees to essentially run any model, regardless of where they come from.
AI Verified source (Sep 8, 2026)

added 1mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
John Villasenor
UCLA professor; Brookings senior fellow

I am this person! Add a sourced quote Report comment

votes Against and says:

Open-weight systems are not inherently less secure than closed systems.
AI Verified source (Aug 10, 2026)

added 1mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Matt Lane
Senior Policy Counsel, Fight for Future

I am this person! Add a sourced quote Report comment

votes Against and says:

Shared access to open-source AI systems is important for protection in a world where AI can be used to hack into electronic systems, like voting systems and benefits infrastructure.
AI Verified source (Aug 10, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Mark Zuckerberg
Meta CEO; Facebook co-founder

I am this person! Add a sourced quote Report comment

votes Against and says:

One of the ways to achieve both individual empowerment and checks and balances is to make sure you put the technology in everybody's hands. I’m personally more worried about centralization than I am about any of the specific risks others are talking ...
more AI Verified source (Aug 10, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Cohesity
Data-security and data-management software company.

I am this person! Add a sourced quote Report comment

votes Against and says:

We're proud to co-sign the Open Weights and American AI Leadership letter alongside 230+ companies, helping enterprises use open AI safely.
AI Verified source (Aug 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
DataRobot
Enterprise AI platform company.

I am this person! Add a sourced quote Report comment

votes Against and says:

DataRobot gives customers the flexibility to run open-weight models on their own terms, which is why we're proud to sign this letter championing an open AI ecosystem.
AI Verified source (Aug 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Resemble AI
AI company developing voice-generation and deepfake-detection models.

I am this person! Add a sourced quote Report comment

votes Against and says:

Releasing these in the open is what keeps us in the research conversation rather than adjacent to it. That research has directly led to our leading multimodal Deepfake Detection Model.
AI Verified source (Aug 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Bria AI
Generative-AI company offering visual-model tools with licensed training data.

I am this person! Add a sourced quote Report comment

votes Against and says:

Bria is proud to sign on to “Open Weights and American AI Leadership,” a coalition statement on why open, inspectable models matter for a healthy, competitive AI ecosystem.
AI Verified source (Aug 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Thinking Machines Lab
AI research and product company publishing work on safe open-weight model releases.

I am this person! Add a sourced quote Report comment

votes Against and says:

Safe open-weight models are public goods, as they put AI development and safety work in many hands and make training choices inspectable.
AI Verified source (Jul 31, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
GitLab
DevSecOps software company; signatory to the Open Weights and American AI Leadership letter.

I am this person! Add a sourced quote Report comment

votes Against and says:

Like policymakers, we want a safe, secure AI ecosystem, and view openness as an important part of achieving it. We support policies that preserve the ability to develop, distribute, and use open weight models subject to focused, risk-based safeguards...
more AI Verified source (Jul 29, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Bill Staples
Chief Executive Officer of GitLab, a DevSecOps software company.

I am this person! Add a sourced quote Report comment

votes Against and says:

Like policymakers, we want a safe, secure AI ecosystem, and view openness as an important part of achieving it.
AI Verified source (Jul 29, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Open Secure AI Alliance
Industry alliance launched in July 2026 to develop open tools and practices for AI safety and security.

I am this person! Add a sourced quote Report comment

votes Against and says:

The alliance used its launch to urge regulators to treat open models as ‘defensive assets, not liabilities,’ warning that blanket restrictions would weaken cyber defenses and concentrate power in a handful of closed AI providers.
AI Verified source (Jul 28, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate

Loading more opinions...

More h/open-source votes

Democratic middle powers should jointly develop frontier AI and share its benefits globally

Require frontier AI labs to conduct nearly all R&D publicly but keep the weights closed

Back