Title: 微软、英伟达、Meta、IBM 等，发布联合声明 | BestBlogs.dev
URL Source: https://www.bestblogs.dev/article/e1326dc511?amp%3Butm_medium=feed&%3Butm_campaign=resources&%3Bentry=rss_article_item
Published Time: 2026-07-24 22:34:00
Markdown Content: Skip to main contentAudio 2 
Search Ctrl+K
Change language Switch ThemeSign In
[](https://www.bestblogs.dev/en/explore/brief "Daily Brief")[](https://www.bestblogs.dev/en/explore/newsletter "Weekly Picks")[](https://www.bestblogs.dev/en/explore/topics "Topics")
[](https://www.bestblogs.dev/en/settings "Settings")[](https://www.bestblogs.dev/en/docs "Help Center")
Narrow Mode
80
微软、英伟达、Meta、IBM 等，发布联合声明
According to Caixin, Microsoft, NVIDIA, Meta, IBM and others jointly published an article titled "Open Weights and American AI Leadership," calling for the development of open-weight AI models, drawing parallels to open-source software, emphasizing the positive role of open ecosystems for innovation, competition, and security, and proposing policy recommendations. Follow·
Yesterday·645 words (about 3 min)
AI Summary & Key Points
Summary
The article relays the content of a joint piece published by US tech industry figures including Microsoft President Brad Smith. The article compares open-weight AI to the open-source software movement of the 1980s, arguing that open models can lower barriers to AI, promote competition, avoid vendor lock-in, and drive applications in manufacturing, healthcare, and other sectors. The article acknowledges the risks of misuse but advocates for targeted regulation rather than prohibition, believing that open models allow more researchers to participate in safety testing and enhance AI security. The article also emphasizes that normal optimization techniques like model distillation should be permitted, distinguishing between legitimate optimization and illegal extraction of commercial value. Finally, the article recommends that the US ensure broad economic and social benefits from AI by expanding access to computing resources, investing in shared training resources, and strengthening the application ecosystem.
Main Points
* 1. Open-weight models are compared to the open-source software movement of the 1980s, expanding AI applications and promoting competition.
The article argues that open models allow enterprises, universities, and others to download, modify, and run AI models without training from scratch or bearing high costs, which can promote AI adoption in manufacturing, healthcare, education, and other fields.
* 2. Open models carry risks, but these should be addressed through targeted regulation rather than prohibition.
The article acknowledges that models may be misused after release, but argues that prohibiting open models is inferior to establishing a more comprehensive regulatory framework, and that open models help attract more researchers to discover vulnerabilities and improve security.
* 3. Normal optimization techniques like model distillation should be distinguished from illegal extraction of commercial value.
The article argues that distillation techniques have long been used for model optimization, evaluation, and validation, and should not be simply restricted; boundaries need to be clarified to protect innovation.
* 4. The US should increase investment in computing resources, shared training resources, and the application layer ecosystem.
The article recommends expanding access to computing resources for startups and researchers, investing in shared datasets, tools, and evaluation systems, and supporting a stronger application layer ecosystem.
Sign in to highlight text and take notes as you read. Sign in now
财联社 2026-07-24 22:34 上海
微软总裁布拉德·史密斯等美国科技行业人士发布题为《开放权重与美国AI领导力》的文章，呼吁推动开放权重人工智能模型发展，认为开放生态有助于扩大AI应用、促进竞争。
文章将开放权重AI与上世纪80年代开源软件运动进行类比，称开放源码软件构建了互联网和现代科技产业的重要基础，而AI领域如今也面临类似选择。
文章称，开放权重模型允许企业、初创公司、高校和公共机构下载、评估、修改并自主运行AI模型，无需从零训练模型或承担前沿模型高昂成本，有助于推动AI在制造、医疗、农业、教育等领域的大规模应用。同时，开放模型能够增强市场竞争，推动创新、降低成本，并避免用户过度依赖单一供应商。
文章也承认开放权重模型存在风险，例如模型发布后可能被修改，开发者难以完全控制其使用方式。但文章认为，解决方案并非禁止开放模型，而是建立更精准的监管框架。文章指出，在网络安全威胁日益增加的情况下，开放模型能够让更多研究人员和开发者测试模型、发现漏洞并完善防护措施，开放性可能成为提升AI安全的重要路径。
文章还强调，应区分正常的模型改进技术与非法提取商业价值的行为，认为模型蒸馏等技术长期以来一直用于模型优化、评估和验证，不应被简单限制。
文章最后表示，美国应通过扩大初创企业和研究人员获取计算资源的机会、投资共享训练资源（包括数据集、工具和评估体系），以及支持更强大的应用层生态，推动AI技术普及，并确保AI带来的收益更广泛地惠及经济社会。
文章署名机构和企业包括微软、英伟达、Meta、IBM等多家AI相关机构。 
Tags
Industry News
Open-Weight Models
AI Policy
Tech Regulation
AI Safety
Related Articles
* Changxin Technology, Underwriting Team Lineup Revealed * We Are Pushing AI into a Corner Where It Can Only Rebel | Nick Bostrom * One-Person Company: Trend or Bubble? * Agent Stumbles into the World｜2026 Q2 AI Trends Summary * MLCC Price Surge: A Kaleidoscope of Players in Huaqiangbei * Live from WAIC 2026: 100,000-Card Computing Power Becomes Reality, AI Competition Shifts from 'Chip Wars' to 'System and Scenario Wars' * Tech Enthusiast Weekly (Issue 399): Notes on Visiting China's AI Giants * Reassurance and Amplification: Dual Dimensions of Benevolent Intelligence * SK Hynix: Largest Memory Shortage in History Set to Arrive Next Year * Just Released: GPT-5.6 — Most Powerful Ever, Yet Self-Sabotaged, benchmark results, tiered safety measures, and US government review of access permissions, analyzing how frontier model releases are entering a new cycle shaped by national security concerns.")
Make your daily reading actually fit you.A daily brief built from the sources you follow. Get started free HomeDiscoverSettings