全球视野 | 2025年慈善领域流行语：自动化筹款人、加速主义、成瘾性智能......

Effective accelerationism, or e/acc, describes a movement that promotes the most rapid and unregulated research and development of A.I. The movement’s adherents argue that those who create the biggest, fastest, most powerful A.I. systems will be able to solve some of the world’s worst problems. They see themselves as an alternative to effective altruism, an evidence-based philanthropic approach that views A.I. as an existential threat.

有效加速主义，简称e/acc，指的是一种主张以最快速度、最少限制地进行人工智能（A.I.）研究和开发的运动。该运动的追随者认为，那些能够创建最大、最快、最强大A.I.系统的人，将能够解决世界上一些最严重的问题。他们将自己视为有效利他主义（effective altruism）的替代方案，后者是一种基于证据的慈善方法，将A.I.视为一种存在性威胁。

A different type of accelerationism inspires white supremacists, one in which they foment divisiveness and polarization to speed up the collapse of existing systems and incite civil war. The people promoting and defending against both versions of accelerationism organize online, in associations, and using nonprofits.

另一种形式的加速主义则激发了白人至上主义者的行动，他们通过煽动分裂和两极分化来加速现有体系的崩溃，并试图引发内战。无论是推动还是抵制这两种加速主义的人，都在网络上、通过协会组织以及非营利机构进行组织和活动。

All A.I. chatbots are working for their maker, not the user. As such, some may incorporate “dark patterns,” or design tactics that manipulate people into certain actions and can get them hooked — similar to YouTube’s “play next” feature. As more people turn to chatbots for conversation and comfort, or to replace human relationships entirely, they need to be on the lookout for A.I. systems with deliberately addicting patterns. Nonprofits considering developing chatbots should err on the side of humanity, and avoid causing technological harm for the people they serve.

所有人工智能聊天机器人都是为其制造者工作，而非其用户。因此，有些聊天机器人可能会采用“黑暗模式”或设计策略，来操纵用户进行某些操作，让用户上瘾——类似于YouTube的“播放下一个”功能。随着越来越多的人开始使用聊天机器人进行对话和寻求慰藉，或者完全取代人际关系，他们需要警惕那些故意让人上瘾的人工智能系统。非营利组织在考虑开发聊天机器人时，应站在人性化的角度，避免给服务对象造成技术伤害。

You knew this was coming. Autonomous fundraisers are A.I. avatars or chatbots that respond quickly, and presumably politely to donors, so they feel cared for. At least until they don’t. This is likely a terrible idea, but nonprofits everywhere will likely want to try it. The big “what if” is whether making fundraising more efficient will also make it more obnoxious and transactional. Inevitably, the bots will negotiate a gift a donor doesn’t actually want to make — or even trick some people into thinking they’re human. Nonprofits should prepare for some spectacular backlash when a donor feels angry or duped.

你知道这迟早会到来。自动化筹款人是一种人工智能化身或聊天机器人，它们能快速回应捐赠人，而且大概会很有礼貌，让捐赠人感觉自己受到了关怀。至少在捐赠人还没有不满之前。这可能是个糟糕的想法，但各地的非营利组织可能都想尝试一下。一个大的“假设”是，是否让筹款更高效，也会让它变得更令人讨厌和交易化。不可避免的是，这些机器人会促成捐赠人做出他们实际上并不愿意做出的捐赠，甚至可能欺骗一些人，让捐赠人误以为它们是人类。当捐赠人感到愤怒或被骗时，非营利组织应该准备好应对强烈的反弹。

Coined by Inside Philanthropy, this term refers to five ultra-high-net-worth donors who primarily give to donor-advised funds. For these billionaires, DAFs function as black boxes because of their lack of disclosure and payout requirements. The article notes that 83 to 99 percent of their giving goes towards DAFs. And as Chronicle of Philanthropy columnist Craig Kennedy has explained, DAF dollars sent to 501(c)(3) organizations can be funneled to 501(c)(4)s that engage in more political activity. For these donors, it seems that the control enabled by DAFs and their anonymity when giving to controversial causes are especially attractive features.

这个词是由《Inside Philanthropy/慈善内幕》创造的，指的是五位超高净值捐赠人，他们主要向捐赠人建议基金捐赠。对这些亿万富翁来说，捐赠人建议基金就像一个黑匣子，因为它们缺乏信息披露和支付要求。《Inside Philanthropy/慈善内幕》的文章指出，这些亿万富翁83%到99%的捐赠都流向了捐赠人建议基金。正如《Chronicle of Philanthropy/慈善纪事报》专栏作家克雷格·肯尼迪所解释的那样，捐赠人建议基金捐给501(c)(3)组织的资金，可能会被流转到从事更多政治活动的501(c)(4)组织。对于这些捐赠人来说，捐赠人建议基金所带来的控制权，以及在向有争议的事业捐赠时的匿名性，似乎是特别有吸引力的特点。

This term refers to a strategy to expand the reach of a program or idea so it spreads like dandelion fluff in the wind. To do so, people aim to ensure their ideas take root in a variety of environments by emphasizing adaptability and resourcefulness. It’s an apt metaphor amid the growing interest in protecting natural habitats, such as the movement to avoid lawn maintenance known as “no mow May.” But the tactics involved in a dandelion strategy can sometimes be vague, and other nonprofits would benefit if those using the metaphor were more specific about how it worked.

这个词指的是一种扩大计划或想法影响范围的策略，旨在使其像蒲公英的种子一样在风中传播。为了实现这一目标，人们试图确保他们的理念能够在多种环境中扎根，强调适应性和应变能力。这是一个恰当的比喻，尤其是在越来越多的人开始关注保护自然栖息地的背景下，比如“五月不修草”运动（即避免修剪草坪的活动）。然而，蒲公英策略所涉及的具体战术有时可能比较模糊，如果使用这个比喻的人能够更具体地说明它是如何发挥作用的，其他非营利组织也会从中受益。

The technology outlet ArsTechnica uses this term to describe the current era in which A.I. fakes take over text, audio, and video, making it impossible to trust anything on the internet. Expanding on the phrase “deep fake,” deep doubt goes beyond one-off examples to encompass all digital media. And it captures both the use and excuse of faked images to explain away reality, a phenomenon known as the “liar’s dividend.” Electoral politics thrives and dies on such information. Civil society, for its part, both creates deep doubt and is a victim of it. Take vaccine misinformation: Major charities have spent millions pushing false science, forcing fact-based organizations to deploy their dollars to fight back.

科技媒体ArsTechnica用这个词来形容当前的时代。在这个时代，人工智能的伪造技术充斥文本、音频和视频，使得人们难以相信互联网上的任何信息。深度怀疑一词扩展了“深度伪造/deep fake”这一词语的内涵，深度怀疑不仅仅局限于个别一次性的伪造案例，而是涵盖了所有数字媒体。它既体现了伪造图像的使用，也包括了为伪造图像辩解的借口，借此掩盖现实，这一现象被称为“骗子的红利/liar’s dividend”。选举政治就是在这样的信息中繁荣和消亡的。就公民社会而言，它既是深度怀疑的制造者，也是受害者。例如疫苗错误信息：大型慈善组织花费了数百万资金传播虚假科学，这迫使以事实为基础的组织动用资金以进行反击。

Corporations gather so many minute data points on our online habits that it’s possible to think of ourselves as having both a physical self and a virtual self created by analyzing personal data. Digital twins can be used currently to predict shopping behavior or provide individualized health guidance. Eventually though, A.I. companies say digital twins will be able to do everything from cure diseases to handle boring tasks, such as attending meetings so we don’t have to. To prepare for that future, civil society needs to consider how to protect the safety and freedom of digital twins, ensuring people can still maintain ownership and control over their data.

企业们收集了我们在线行为的海量细节数据，以至于我们可以把自己看作既有一个物理自我，又有一个通过分析个人数据生成的虚拟自我。目前，数字孪生可以用来预测购物行为或提供个性化的健康指导。然而，人工智能公司表示，数字孪生最终将能做任何事，从治愈疾病到处理无聊的任务（例如代替我们参加会议）等一切事情。为了迎接这个未来，公民社会需要考虑如何保护数字孪生的安全与自由，确保人们依然能够保持对自己数据的所有权和控制权。

Just like spam has long filled email inboxes, A.I.-generated junk, or slop, now fills the web. While spam can be blocked at an inbox’s metaphorical door, slop can’t be stopped. It’s everywhere on the internet: an image of Jesus with shrimp for limbs on Facebook. Dangerously inaccurate foraging books on Amazon that may lead people to eat poisonous mushrooms. The more junk on the web, the harder it is to find accurate information that good actors in civil society are sharing. The challenge for any organization is how to be seen and heard amidst all the slop.

就像垃圾邮件长期以来充斥着电子邮件收件箱一样，人工智能生成的垃圾内容现在充斥着互联网。虽然垃圾邮件可以在收件箱的“象征性门口”被拦截，但垃圾内容却无法被阻止。它无处不在：比如在Facebook上一张四肢是虾的耶稣图片；或者在亚马逊上有误导性的觅食书籍，可能导致人们误食有毒蘑菇。互联网垃圾内容越多，就越难找到民间社会中的积极参与者所分享的准确信息。任何组织面临的挑战是，在这片垃圾信息的海洋中，如何让自己被看见并发出声音。

Where once organizations focused on making cogent, one-off arguments to persuade others, now they socialize an idea by trying to build support for a concept over time, introducing the information slowly and in digestible pieces. This might involve presenting the idea to people one-on-one, getting their feedback, and tweaking the approach where needed to ensure that eventually, the entire organization is on board. People one level below the final decision makers often speak of “socializing the idea, the proposal, the strategy” among their bosses.

过去，组织通常专注于通过有力的、一次性的论据来说服他人。而现在他们通过“社交化”理念来尝试建立对某一概念的支持，以逐步的、易于消化的方式引入信息。这可能包括一对一地向人们展示理念，获取反馈，并在需要时调整方法，确保最终整个组织都能达成一致。比最终决策者低一级的人，常常提到在他的上级之中“社交化想法、提案或策略”。

In the country’s Alice-through-the-looking-glass politics, to weaponize something is to turn a legitimate issue or concern into a tool for inflicting political damage. Recently, for example, Republican lawmakers accused misinformation researchers of censorship, even though they were the ones doing the censoring by issuing subpoenas.

在美国的“爱丽丝镜中奇遇记”式（译者注：指该国的政治环境像爱丽丝在镜中世界中的经历一样荒诞、颠倒）的政治生态中，武器化是指将正当的议题或关切转化为制造政治杀伤的工具。最近的典型案例是：共和党议员指控不实信息研究人员在实施审查，尽管这些议员才是通过签发传票来进行审查的人。

At least one Congressional subcommittee focuses on the weaponization of government power. Leonard Leo, the right-wing political activist with a billion dollar-plus bank account, called on his grantees to “weaponize” their work two months before the 2024 U.S. presidential election. In the dark money sandbox of politics and charity that Leo inhabits, real issues that matter to real people, from marriage equality to health-care access, are cultural divisions to be exploited. Expect this behavior and language to creep further into the charitable sector in the coming year. 

至少有一个国会小组委员会关注政府权力的武器化。右翼政治活动家、拥有超过十亿美元资产的莱昂纳德·利奥，在2024年美国总统选举前两个月，呼吁他的受资助人“武器化”他们的工作。在利奥所处的政治与慈善的黑金沙盒里，从婚姻平等到医疗保健，这些与普通人息息相关的现实问题都成为了可以利用的文化分歧。预计这种行为和语言将在未来一年进一步渗透到慈善领域。