美国中期选举将于周二举行。选举结束后，OpenAI将立即展开一场密集的政策倡议和公关行动，以争取适应华盛顿新的政治格局。其中一项举措是重新包装数据中心的概念，将其称为“人工智能算力园区”（AI compute campuses）。
在过去两年里，人工智能企业一直与白宫保持着友好关系。但周二过后，随着民主党很可能赢得众议院控制权，这些企业将面临一个态度更加强硬的华盛顿。与此同时，OpenAI还希望影响2028年美国总统大选的人工智能政策辩论。届时，潜在总统候选人的阵营将从周三开始逐渐形成。
OpenAI首席全球事务官克里斯·莱恩（Chris Lehane）告诉我们：“你可以选择追逐时代潮流，也可以尝试通过思想的力量影响时代潮流。”他将周二的选举称为“从一个时代迈向下一个时代的选举转折点”。
自8月以来，莱恩及其团队一直在储备政策提案，其中既有全新构想，也有经过重新包装的旧有方案，旨在强化外界对人工智能政策问题的紧迫感，并引导相关辩论的方向。
这一系列行动将在美国国会的“跛脚鸭会期”（lame-duck session）率先展开。这是指选举结束后、直到新一届众议员和参议员于1月3日宣誓就职前的过渡时期。虽然这一阶段的政治局面往往较为尴尬，但仍可能产生重大政策影响。今年的跛脚鸭会期将于11月9日开始。
大局：科技企业纷纷为民主党可能掀起的“蓝色浪潮”做准备
其他科技企业也在为民主党可能取得大胜的局面提前布局。
Meta正在采取措施，加强与民主党的联系。预计其他大型人工智能企业和科技公司也将采取类似行动。
那些与特朗普总统关系密切的科技巨头，将成为民主党重点针对的对象。民主党希望加强对科技行业的审查和监管压力。
OpenAI的相关计划此次首次得到披露，其核心主题均围绕着一个概念：控制权。OpenAI试图传递这样的信息：人工智能的发展是为了服务人类，而不是取代人类。
具体提案包括以下几个方面：
一、设置人工智能发展的“速度限制”
在此前提案的基础上，OpenAI计划公布新的政策构想，要求对能力最强大的人工智能系统实施强制性安全保障措施。
这些措施相当于为人工智能设置“速度限制”，确保人类始终能够掌握技术发展的方向盘。OpenAI还将加大在国会推动相关立法的力度。
莱恩表示：
“你们将看到，我们越来越多地讨论‘速度限制’这一概念。”
“我们相信，人类应当始终掌握这项技术的控制权。人工智能应当由人类创造、为人类服务，而人们应当对这项技术如何塑造自己的生活和社区拥有更大的决定权。”
二、数据中心：重新塑造人工智能基础设施的形象
面对美国各地出现的、跨越两党界限的数据中心建设反对浪潮，OpenAI希望通过提出新的社区合作协议和承诺，为未来的数据中心项目扫清障碍。
此外，该公司还计划提出州级立法建议，试图建立新的行业标准，确保当地居民能够切实参与决定项目的建设方式，以及社区能够从中获得哪些收益。
在一项有意引发讨论的品牌重塑行动中，OpenAI将主张，其建设的基础设施与传统数据中心存在本质区别，因此应被称为“人工智能算力园区”（AI compute campuses）。OpenAI认为，这类园区能够推动所在地社区的经济转型。
OpenAI特别提到了其参与开发的一个大型数据中心项目。该项目位于俄亥俄州阿巴拉契亚地区的派克县（Pike County），预计将为当地创造数千个就业岗位。
莱恩表示：
“我们认为，这样的讨论是积极而健康的。这正是民主制度发挥作用的体现。”
“社区应当要求企业承担责任、接受监督，确保当地真正从这些项目中获得具有变革意义的收益。”
三、经济机会：让更多人从人工智能中受益
OpenAI正在制定一系列政策构想，希望让更多人认识到，人工智能能够帮助他们改善日常生活、提高工作能力，而不仅仅是一种潜在威胁。
相关措施包括：帮助劳动者应对人工智能带来的就业冲击；扩大教育和职业培训资源的覆盖范围；降低创业和企业发展的门槛。
深层含义：在两种极端立场之间寻找“第三条道路”
人工智能政策辩论往往被简化为两种相互对立的立场：一方是担忧人工智能可能带来灾难性后果的“末日论者”（doomers），另一方则是主张加速技术发展的“加速主义者”（accelerationists）。
OpenAI正试图打破这种人为制造的二元对立，提出一条以人为中心的“第三条道路”。
今年9月，OpenAI首席执行官萨姆·奥尔特曼（Sam Altman）在联合国安理会发表讲话时，呼吁采取一种能够催生“创造力与发现的新文艺复兴”的人工智能发展路径。
接下来值得关注：OpenAI试图影响2028年美国总统大选
OpenAI的政策示好部分面向2028年美国总统大选的潜在候选人。这些政治人物预计将从周三开始，越来越积极地公开阐述自己的施政构想。
莱恩表示：“人工智能将成为2028年大选的核心议题。”
他将此次中期选举形容为“2028年大选的一段电影预告片”。
莱恩进一步表示：
“我认为，光说一句‘人工智能不好’，已经不足以赢得公众支持。”
“公众明白，这是一场远比简单的好坏判断更加复杂、更加微妙的讨论。他们将寻找属于这个时代的罗斯福或林肯——一个真正愿意带领他们走过这一转型历程的领导者。”
莱恩认为，总统候选人应当思考以下问题：
“我们该如何开始思考信任问题？该如何看待现代社会中普遍存在的巨大焦虑？又该如何理解这样一个事实：人们对当前的社会契约状况，以及美国梦的现实处境，感到极度不满？”
OpenAI将持续推动华盛顿加快人工智能监管行动
OpenAI未来几个月的计划，只是该公司试图增强华盛顿政策制定紧迫感、推动整个行业更加重视前沿人工智能安全的开端。
需要记住的是：人工智能模型的能力每天都在增强，而要将其控制在合理范围内，也正变得越来越困难。
莱恩表示：
“我们的政策优先事项，尤其是我们在前沿人工智能安全问题上强调的紧迫性，都是为了回应技术能力快速进步所带来的时代需求。”
“随着新一届国会开始履职，我们预计这些努力还将继续。”
本文由 AI图鉴自动抓取并翻译，版权归原作者所有；如需查看原文请点下方链接。
United States midterm elections will be held on Tuesday. Immediately after the elections, Open AI will launch an intensive policy initiative and public relations initiative to adapt to the new political landscape in Washington. One such initiative is the repackaging of the data centre concept, known as the “AI computer parks”.
Over the past two years, artificial intelligence enterprises have maintained friendly relations with the White House. But after Tuesday, as the Democrats are likely to win control of the House, they will face a stronger attitude in Washington. At the same time, Open AI hopes to influence the artificially intelligent policy debate in the 2028 presidential election. At that time, the formation of a camp of potential presidential candidates will begin on Wednesday.
Open AI Chief Global Affairs Officer Chris Lehane told us, "You can choose to pursue the tide of times, or you can try to influence the tide of times through the power of ideas." He called Tuesday's elections “an electoral turning point from one era to the next”.
Since August, Ryan and his team have been preparing a proposal for a policy that includes both a new vision and a repackaged old programme aimed at strengthening external sense of urgency on artificial intelligence policy issues and guiding the debate.
This series of actions will be spearheaded at the “lame-duck session” in the United States Congress. This refers to the transitional period following the election until the swearing-in of the new House of Representatives and Senators on 3 January. While the political situation at this stage is often more embarrassing, it may still have significant policy implications. This year's lame duck session will begin on 9 November.
The big picture: technology companies are preparing for the "Blue Wave" that the Democrats might launch.
Other technology firms are also pre-positioning the Democratic Party’s potential victory.
Meta is taking steps to strengthen ties with the Democratic Party. Similar action is expected to be taken by other large artificial intelligence enterprises and technology companies.
Those technocrats close to President Trump will be the focus of the Democratic Party. The Democratic Party wants to increase censorship and regulatory pressure on the science and technology industry.
Open AI's plan was first disclosed, and its core theme revolved around one concept: control. Open AI tries to convey the message that artificial intelligence is developed to serve humanity, not to replace it.
Specific proposals include the following:
I. Setting “speed limits” for artificial intelligence development
On the basis of previous proposals, Open AI plans to publish a new policy concept requiring mandatory safeguards for the most capable artificial intelligence systems.
These measures are tantamount to setting “speed limits” for artificial intelligence to ensure that humankind can always master the wheel of technological development. Open AI will also step up its efforts to promote relevant legislation in Congress.
Ryan says:
“You will see that we are increasingly discussing the concept of `speed limits'.”
“We believe that humankind should always have control over this technology. Artificial intelligence should be created by, and at the service of, human beings, and people should have greater discretion in how this technology shapes their own lives and communities.”
Data centres: reshaping the image of artificial intelligence infrastructure
In the face of a wave of opposition across the United States to the construction of data centres across bipartisan boundaries, OpenAi hopes to clear the way for future data centre projects by proposing new community cooperation agreements and commitments.
In addition, the company plans to propose state-level legislation in an attempt to establish new industry standards to ensure the effective participation of the local population in determining the manner in which projects are built and the benefits that the community can derive from them.
In a re-branding exercise intended to stimulate discussion, Open AI will argue that the infrastructure it has built is fundamentally different from the traditional data centres and should therefore be referred to as the “AI computer campus”. Open AI believes that such parks can contribute to the economic transformation of local communities.
Open AI referred in particular to its participation in the development of a large data centre project. The project, located in Pike County, Appalachia District, Ohio, is expected to create thousands of local jobs.
Ryan says:
“We believe that such discussions are positive and healthy. This is the manifestation of the functioning of democracy.”
“Communities should hold businesses accountable and subject to supervision to ensure that local gains from these projects are truly transformative.”
III. Economic opportunities: increasing the number of people benefiting from artificial intelligence
Open AI is developing a range of policy ideas that will make people more aware that artificial intelligence can help them improve their daily lives and improve their ability to work, not just as a potential threat.
Measures include: helping workers to cope with the employment shocks of artificial intelligence; expanding the coverage of educational and vocational training resources; and lowering the threshold for entrepreneurship and enterprise development.
Deep meaning: the search for a “third path” between the two extreme positions
Artificial intelligence policy debates are often reduced to two opposing positions: “end-of-lifeists” who fear the catastrophic consequences of artificial intelligence, and “accelerators” who advocate accelerated technological development.
Open AI is trying to break this artificial dichotomy by proposing a human-centred “third path”.
In September this year, in his address to the United Nations Security Council, CEO Open AI Sam Altman called for an artificially intelligent development path capable of generating a “new Renaissance of creativity and discovery”.
The following is a matter of concern: Open AI is trying to influence the 2028 presidential election in the United States.
Open AI’s policy is aimed in part at potential candidates for the 2028 United States presidential election. These politicians are expected to be increasingly active in articulating their governance visions publicly, starting on Wednesday.
“Advisory intelligence will be at the heart of the 2028 general election.”
He described this mid-term election as “a film preview of the 2028 general elections”.
Ryan further stated:
“I don't think it's enough to say that `a bad artificial intelligence' is enough to win public support.”
“The public understands that this is a much more complex and delicate discussion than a simple good or bad judgement. They will look for Roosevelt or Lincoln, who belongs to this era — a leader who is truly willing to lead them through this transition.”
Ryan believes that presidential candidates should reflect on the following issues:
“How should we start thinking about trust? How should the great anxiety prevailing in modern society be viewed? How can one understand the fact that people are extremely dissatisfied with the current state of the social contract and the reality of the American dream?”
Open AI will continue to push Washington to speed up AI regulatory action.
Open AI's plans for the coming months are only the beginning of an attempt by the company to increase the sense of urgency of Washington's policy development and promote greater industry-wide focus on front-line artificial intelligence security.
It is important to remember that the capability of artificial intelligence models is increasing every day and that it is becoming increasingly difficult to keep them within reasonable limits.
Ryan says:
“Our policy priorities, and especially the urgency that we emphasize on front-line issues of artificial intelligence security, respond to the demands of the times brought about by rapid technological advances.”
“We expect these efforts to continue as the new National Assembly begins its functions.”
Auto-fetched and translated by AI Atlas. All rights belong to the original author.