Mashable 07月25日 04:48
Trump sides with Big Tech in AI copyright battle — "China’s not doing it"
index_new5.html
../../../zaker_core/zaker_tpl_static/wap/tpl_guoji1.html

 

美国总统特朗普在AI峰会上公布了其AI行动计划,并明确表示支持科技公司在AI版权争议中的立场。他主张对AI和知识产权规则采取“常识性应用”,认为不能要求AI开发者为学习过程中接触到的所有公开信息付费,否则将阻碍AI项目的成功。这一表态与Meta、OpenAI等大型科技公司在AI模型训练和版权侵犯问题上的观点一致,即认为使用公开数据进行模型训练属于合理使用。然而,此番言论也引发了关于知识付费和版权保护的讨论,有法律专家指出,学习知识并不意味着可以规避版权法,AI公司仍需为使用受版权保护的材料进行许可或支付赔偿。文章还提到,白宫发布的AI行动计划中虽未直接提及版权问题,但特朗普的言论预示了政府在此问题上的倾向性,并指出其立场可能与美国版权局及国会提出的相关法案存在差异。

🎯 **特朗普政府支持科技公司在AI版权争议中立场**:美国总统特朗普在其AI行动计划中,明确表达了对科技公司在AI版权问题上立场的支持,认为AI开发者不应为学习过程中接触到的公开信息付费,否则将阻碍AI发展。他强调,“你不能期望有一个成功的AI项目,当你阅读或学习的每一篇文章、每一本书,你都应该付费。”这反映了他对AI训练数据获取方式的看法,并将其与知识获取的常识性原则联系起来。

⚖️ **AI公司主张合理使用,版权方指控侵权**:以Meta和OpenAI为代表的科技公司认为,使用公开可用的数据训练AI模型属于“合理使用”(fair use)的范畴,并主张这对于保持美国的AI竞争力乃至国家安全至关重要。然而,包括作者、音乐家和艺术家在内的版权所有者则提起诉讼,指控AI公司使用其作品(部分为盗版)作为训练数据构成版权侵犯,认为AI生成器如Midjourney是“盗版无底洞”。

📚 **知识获取与版权付费的辩论**:文章指出,特朗普关于“不为获取知识付费”的说法是一种常见的论点,但许多版权律师对此持不同意见。他们认为,学习知识并不代表可以违反版权法,AI公司需要为使用大量受版权保护的材料进行许可,或在侵权时支付赔偿。有法律专家明确表示,“人们付费学习,他们必须购买书籍副本才能学习。”这揭示了AI训练数据获取与传统知识获取方式在版权法律解释上的冲突。

📜 **政策动向与潜在分歧**:尽管特朗普的言论预示了政府对科技公司有利的立场,但白宫发布的AI行动计划并未明确提及版权或合理使用问题。此外,特朗普的立场可能与美国版权局的初步结论以及国会提出的禁止AI公司使用受版权作品进行训练的法案存在潜在分歧。文章也提及,特朗普本人未来也可能因AI公司训练其节目内容而不付费而改变立场。

President Donald Trump is siding with tech companies in the hotly contested AI copyright battle.

On Wednesday, the president outlined his AI Action Plan in remarks at the AI Summit, an event hosted by the All-In podcast and the Hill and Valley Forum. In announcing the plan, Trump called for a "common-sense application" of AI and intellectual property rules (starting at minute 19:48 in the video).

"You can't be expected to have a successful AI program when every single article, book, or anything else that you've read or studied, you're supposed to pay for," the president said. "When a person reads a book or an article, you've gained great knowledge. That does not mean that you're violating copyright laws or have to make deals with every content provider." Notably, he also said, "China’s not doing it."

The rhetoric reflects the stance big tech companies like Meta and OpenAI have taken when it comes to AI model training and copyright infringement.

AI companies say copyright law is slowing them down

A fierce legal debate is raging that pits copyright holders against AI companies. Authors, musicians, and artists have sued OpenAI, Meta, Anthropic, and others, claiming the use of their creative works (some of which were pirated) as training data is copyright infringement. For instance, in its suit against Midjourney, Disney called the AI image generator "a bottomless pit of plagiarism." However, AI companies claim they are protected by the fair use doctrine, which allows the unlicensed use of copyrighted works in specific circumstances.

In its March recommendations for development of an AI Action Plan, Meta said, "the Administration should clarify that the use of publicly available data to train models is unequivocally fair use." OpenAI similarly argued that "applying the fair use doctrine to AI is not only a matter of American competitiveness — it’s a matter of national security," in their submitted recommendations. Companies like OpenAI have claimed they need the green light to use copyrighted works without consent to win the AI race against China.

As part of its AI Action Plan, the White House also issued a trio of executive orders about AI, which Trump signed after giving his remarks. There was no explicit mention of copyright issues or fair use in these orders, but Trump's comments clearly signal that the administration is siding with big tech on this issue. The action plan also calls for easing regulations on tech companies in general. "To maintain global leadership in AI, America’s private sector must be unencumbered by bureaucratic red tape," said the AI Action Plan.

Fact check: You do have to pay for knowledge

Trump's comments about not having to pay for gaining knowledge is actually a common argument. Tech companies have claimed that because their AI models are generally learning from the data, they are not subject to copyright laws.

However, many copyright lawyers disagree.

"When the president raises the difficulty of paying for every 'single article, book, or anything else,'that’s legit," said Justin Hughes, a law professor at Loyola Law School. "But nothing in that is inconsistent with the AI companies licensing large repositories of copyrighted materials – as they are doing, AI companies reaching settlements with certified classes of copyright owners in litigation – as they should, and AI companies paying exemplary damages when they rapaciously copy stuff without permission."

In a previous interview with Mashable about the Kadrey v. Meta case, Hughes also pointed out that "people pay to learn and they have to buy copies of books to learn."

Dozens of AI copyright cases are making their way through the courts right now. AI blog ChatGPT Is Eating the World counts 47 active cases. Recently, Anthropic and Meta both won their fair use rulings, but the judge presiding over the Meta case had this to say: "As should now be clear, this ruling does not stand for the proposition that Meta’s use of copyrighted materials to train its language models is lawful," and even postulated that "it will be illegal to copy copyright-protected works to train generative AI models without permission."

It's unclear how Trump's copyright remarks will turn into concrete action. But he seems to be out of sync with the U.S. Copyright Office's position on AI training and fair use, according to a pre-publication version of its AI Copyright Report. That report concluded that fair use likely doesn't provide a blanket exception for AI companies using copyrighted works. In Congress, Senators Josh Hawley and Richard Blumenthal recently introduced a bipartisan bill that would ban AI companies from training on copyrighted works and allow people to sue tech companies for the use of their personal data for model training.

And then there's the possibility that Trump's stance might change, as Hughes theorized. "I think the president may feel very differently if and when he learns about AI companies training on episodes of The Apprentice without paying a dime."


Disclosure: Ziff Davis, Mashable’s parent company, in April filed a lawsuit against OpenAI, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.

Fish AI Reader

Fish AI Reader

AI辅助创作,多种专业模板,深度分析,高质量内容生成。从观点提取到深度思考,FishAI为您提供全方位的创作支持。新版本引入自定义参数,让您的创作更加个性化和精准。

FishAI

FishAI

鱼阅,AI 时代的下一个智能信息助手,助你摆脱信息焦虑

联系邮箱 441953276@qq.com

相关标签

特朗普 AI版权 科技公司 合理使用 知识产权
相关文章