谁来制定人工智能的规则?
人工智能能为我的孩子带来更美好的未来吗?
人工智能会让世界变得更危险吗?
人工智能能帮助科学家治愈疾病吗?
人们对人工智能有很多尖锐的问题。回应这些问题正是我们的职责所在。
许多人对人工智能持积极态度。他们每天都在使用它,并看到了它的潜力:让我们的工作和生活不再那么繁重,改变我们学习的方式,帮助加速科学和技术进步,创造新的繁荣源泉,以及解决我们面临的一些最大的医疗和社会问题。
但也有许多人抱有严重的担忧。一些人担心人工智能可能导致失业。一些人担心它可能会贬低创造性工作的价值。另一些人则关注人的自主性:人工智能可能会影响我们独立思考、建立人际关系以及在生活中寻找意义的能力。许多人担心,如果人工智能的能力落入不法之徒手中会意味着什么——并质疑其益处是否大于代价。
在下面的视频中,您可以听到我们采访过的人们所表达的这些希望与担忧。
Anthropic 是一家公益公司——我们的使命是确保先进 AI 模型的益处,并降低其风险。例如,这一公益使命促使我们投资于 AI 安全防护措施以减少滥用的风险,研究 AI 模型的行为和内部运作以帮助我们将它们与有益目标对齐,免费向科学家提供我们的 AI 模型,并启动了一个奖学金项目,将处于职业生涯早期的 Claude 用户与非营利组织配对,以帮助扩大 AI 的益处。
要真正以公共利益为出发点行事,我们需要充分了解公众对这项新技术的希望与担忧。作为我们关于“尖锐问题”新计划的一部分,我们正在采取进一步措施,了解人们对各种 AI 相关话题的看法和疑问,并透明地记录我们在实现公益目标方面的进展。
我们已经在为此奠定基础,向数千人询问了他们对人工智能的看法:
- 我们启动了“Anthropic 公共记录”——这是一项公开调查,首轮已邀请 52,000 名美国人阐述他们对这项技术最大的期望与担忧;
- 我们通过“Anthropic 访谈员”工具,对来自 159 个国家、使用 70 种语言的 81,000 名 Claude 用户进行了调查;
- 我们举办了数十场线下焦点小组讨论,并召集了那些其工作与传统与 AI 所引发问题息息相关的群体进行座谈;
- 我们还通过匿名化的真实世界数据,持续研究 Claude 的使用情况。
我们创立了 Anthropic 研究所——这是公司内部的一项研究计划,旨在应对 AI 将给社会带来的最重大挑战——并且自公司成立初期以来,我们的“长期利益信托”就一直提供公正的监督,评估我们在推进公益使命方面的成效。
现在,我们明确邀请您向我们提出关于 AI 的最棘手问题:关于 AI 对就业、社会和家庭影响的问题;关于我们如何实现 AI 在科学和医学领域巨大潜力的问题;以及关于人类历史上最强大的技术之一将引领我们走向何方的问题。
作为回应,我们将公开追踪并报告为解决这些问题所采取的具体行动——同时也会坦诚说明我们在哪些方面可能未能达到既定目标。
要查看其他人提出的问题,并分享您自己的问题,请访问我们的“棘手问题”网站。
推出 Claude for Teachers
Anthropic 承诺向加拿大 AI 研究提供 1000 万美元资助
Who decides the rules for AI?
Can AI give my children a better future?
Does AI make the world a more dangerous place?
Can AI help scientists cure diseases?
People have a lot of hard questions about AI. It’s our job to address them.
Many people are positively disposed to AI. They already use it every day, and they see its potential for making our work and our lives less laborious, for changing the way we learn, for helping speed up scientific and technological progress, for creating new sources of prosperity, and for solving some of the biggest medical and social problems we face.
But many also hold serious concerns. Some are worried about AI’s potential contribution to job loss. Some fear that it could devalue creative work. Others are concerned about human agency: that AI might affect our ability to think for ourselves, to make human connections, and to have meaning in our lives. Many are concerned about what it means if AI’s capabilities fall into the wrong hands—and wonder whether the benefits outweigh the costs.
In the film below, you can hear some of these hopes and concerns from people we’ve spoken with.
Anthropic is a Public Benefit Corporation—it’s our mission to secure the benefits of advanced AI models and mitigate their risks. That public benefit mission has led us, for example, to invest in AI safeguards to reduce the risk of misuse, to research the behavior and inner workings of AI models to help us align them to beneficial goals, to give our AI models for free to scientists, and to launch a fellowship program that pairs early-career Claude users with nonprofits to help extend the benefits of AI.
To truly act in the public interest, we need to fully understand the hopes and concerns the public has about this new technology. As part of our new initiative on hard questions, we’re taking further steps to understand people’s views and questions on a variety of AI-related topics, and transparently chart our progress toward our public benefit goals.
We’ve already been laying the groundwork, asking thousands of people for their views on AI:
- We’ve begun the Anthropic Public Record—a public survey that, in its first round, asked 52,000 Americans to set out their biggest hopes and concerns about this technology;
- We surveyed 81,000 Claude users across 159 countries and 70 languages through our Anthropic Interviewer;
- We conducted dozens of in-person focus groups and convened sessions with groups whose work and traditions bear on the questions raised by AI; and
- We’ve been studying the use of Claude through anonymized, real-world data.
We created the Anthropic Institute—a research effort within the company with the goal of confronting the most significant challenges that AI will pose to society—and since early in our company’s history our Long-Term Benefit Trust has provided impartial oversight on how effectively we’re advancing our public benefit mission.
Now, we’re explicitly asking you to send us your hardest questions on AI: questions about AI’s effects on jobs, society, and families; questions about how we achieve the dramatic potential of AI for science and medicine; questions about where one of the most powerful technologies in human history can take us next.
In return, we’ll publicly track and report the specific actions we’re taking to address those questions—and we’ll be clear about the ways in which we might fall short of our stated goals.
To see some of the questions others are asking, and to share your own, visit our hard questions website.