我们能否修复互联网？
内容总结：
当前互联网正面临算法成瘾、数据剥削与虚假信息泛滥等多重挑战。三位业界权威人士在新著中提出了截然不同的治理方案，引发广泛讨论。
哥伦比亚大学教授蒂姆·吴在《提取时代》中指出，科技巨头已从服务用户转向榨取用户，其通过制造“便利陷阱”形成行业垄断。他主张运用反垄断法等现有法律工具进行拆分，并援引历史上成功拆分AT&T和IBM的案例佐证其可行性。不过近年针对谷歌和微软的反垄断诉讼成效不彰，显示传统法律手段在平台时代面临局限。
前Meta高管尼克·克莱格则在《如何拯救互联网》中反对拆分巨头，认为这会影响用户体验。他主张通过内容监管和企业自律实现改革，建议建立全球数据流动条约取代各国单独立法。但因其Meta背景及对平台责任的淡化立场，这些主张被质疑缺乏说服力。
万维网发明者蒂姆·伯纳斯-李在回忆录中提出技术解决方案“Solid”系统，试图通过去中心化数据存储实现用户数据自主权。该方案虽具创新性，但如何确保敏感数据安全、协调人工智能伦理等问题仍需完善。
尽管三位专家方案各异，但都认同加强用户数据控制权、提升平台透明度与问责机制是改革方向。在现有法律框架见效缓慢的背景下，克莱格提出的由美国主导达成全球数字协议的建议，可能成为当前最具操作性的过渡方案。互联网治理道阻且长，但追求更安全、公平的数字环境已成为社会共识。
中文翻译：
我们能否修复互联网？
从政府监管到用户责任，三本新书提出了形形色色的改良方案。
从成瘾算法到剥削性应用，从数据挖掘到虚假信息，当今互联网已成为危机四伏之地。三位重量级人物——"网络中立"理论奠基人、Meta前高管、万维网发明者——在其著作中提出了颠覆性的修复方案。但这些精英真是担此重任的合适人选吗？尽管他们展现了信念甚至独创性，其解决方案仍存在认知盲区。
在《榨取时代：科技平台如何征服经济并威胁未来繁荣》中，蒂姆·吴指出少数平台公司权力过度集中，必须予以拆分。这位率先提出"自由网络需平等对待所有数据流"原则的哥伦比亚大学教授认为，现有法律机制特别是反垄断法，是实现该目标的最佳途径。
蒂姆·吴结合经济理论与数字发展史，揭示了平台如何从服务用户转向榨取用户。他指出人类对其权力的认知缺失助长了垄断态势，而便利性正是平台束缚用户的利器。"人类规避不必要痛苦与麻烦的欲望"，他写道，"或许是世间最强大的力量"。他以谷歌和苹果的"生态系统"为例，展现无缝衔接服务如何令用户产生依赖。对蒂姆·吴而言，这本身并非坏事——亚马逊在娱乐流媒体、在线购物和生活管理方面带来的便捷确实惠及用户。但当亚马逊、苹果和字母表等巨头以便利性征服海量用户并扼杀竞争时，这种"行业统治"就该重新审视。
蒂姆·吴倡导的措施最具实操性：依托现有法律框架与经济政策，运用联邦反垄断法、限制企业服务收费的公用事业上限，以及禁止企业涉足特定行业的"业务范围"限制。
这位学者强调反垄断条款是有效武器，并列举两起成功案例：1960年代美国政府诉IBM案催生了苹果与微软等企业，1982年AT&T分拆案则带来更多市场竞争。但历史经验能否适用于平台时代？2025年谷歌反垄断案中，法官未支持司法部分拆Chrome浏览器的提议；2001年微软反垄断案也未能成功分拆其浏览器业务——这些均暴露了法律手段的局限性。值得注意的是，蒂姆·吴在论证时并未提及微软案例。
尼克·克莱格——Meta前全球事务总裁、英国前副首相——则持截然相反的立场：分拆科技巨头实属误入歧途，只会损害用户体验。在《如何拯救互联网：AI与政治冲突时代的全球连接威胁》中，他承认科技巨头垄断网络，但认为反垄断等惩罚性法律措施徒劳无益，应通过规范社交媒体内容等监管手段规避。（值得玩味的是，Meta正面临其收购Instagram和WhatsApp是否构成垄断的诉讼。）
克莱格同时主张硅谷应主动改革，强调社交媒体网络"公开账本"、与用户分享决策权比诉诸法律更能重建平衡。但这位曾与扎克伯格共事却未能推动变革的政客难免令人存疑。更可疑的是其著作对历史的选择性呈现：虽提及剑桥分析公司2016年窃取Facebook数据等丑闻，却回避讨论Facebook导致网络割裂的自身责任。
克莱格认为反垄断分拆将阻碍创新，"完全无视用户从网络效应中获得的利益"。他写道用户留守这些超级平台是因能找到"大部分所需"，无论是社交媒体的好友与内容，还是亚马逊和易趣的廉价商品。蒂姆·吴或会认同此点，但必然反对其维持现状有益用户的论断。"传统反垄断逻辑已然失效"，克莱格坚持认为，有限度的监管才能在改善用户体验的同时降低科技巨头危害。
拥有立法者与企业高管双重视角的克莱格，既参与过制定科技法规的英国政府工作，又协助Meta应对各国监管要求。他抱怨硅谷为遵守全球不同规则疲于奔命，但这些掌握庞大资源的企业岂会无力应对？鉴于Meta曾在印度因违反网络中立原则被禁售免费基础服务，克莱格的抱怨更显可疑。他主张真正要务并非各国单独立法，而是建立保护数据跨境流动的全球条约。
这位前Meta高管的改革方案——不出所料——并非分拆巨头，而是推动其实现"彻底透明"。他要求平台让用户更多参与治理（如借助Facebook社群论坛模式优化产品），并赋予用户对个人数据及其使用方式的实质控制权。
在此点上，克莱格与万维网发明者蒂姆·伯纳斯-李不谋而合。后者在回忆录兼宣言《献给众生：万维网未竟之路》中承认，他最初设想的开源、协作、完全去中心化网络与当今现实相去甚远。维基百科作为"最接近我理想中网络形态的范例"，成为其原始蓝图的仅存火种。
伯纳斯-李将数据控制权交还用户的核心方案，是他参与开发的通用数据"舱"Solid（"社交关联数据"缩写）。这个源自麻省理工学院的系统提供中央存储站，可管理从信用卡信息、健康记录到社交评论等全维度数据。"与其让数据散落各方，不如将完整数字足迹存入私人仓库"，他写道。
在数据窃取猖獗的时代，Solid宛若灵丹妙药。让用户掌控自身数据并知晓"被生成何种数据"确实诱人。但人们难免担忧——例如将健康记录与智能手表心率数据合并的风险。尽管伯纳斯-李承诺用户控制与去中心化，经期追踪应用滥用数据等丑闻仍令人心有余悸。
伯纳斯-李认为Solid能节省时间改善网络生活，抱怨如今在不同航司App反复填写信息的繁琐。Solid的数字保险库可整合疫苗接种记录与信用卡交易等数据，人工智能更能助其发挥价值（如关联膳食计划与杂货账单）。不过他对AI与Solid协同造福用户的展望虽乐观，却未明确说明如何确保聊天机器人敏感数据的安全性。
除青少年与社交媒体算法领域外，伯纳斯-李普遍反对网络监管。他坚信用户拥有数据自主权，期待Solid能将网络从"榨取式"发展轨道矫正归来。
三位作者的改革方案中，蒂姆·吴的主张已初见成效：谷歌被迫通过数据共享给予竞争对手便利，其系统在新产品技术中的应用也受到限制。但在美国当前政治气候下，反垄断法会否持续约束科技巨头？克莱格在限制各国立法方面或能如愿——特朗普已确认将用关税惩罚针对美国科技公司的立法国家。鉴于其政府姿态，美国加强对科技巨头的监管可能性渺茫。事实上，社交媒体正显现扩张态势（如Meta在特朗普胜选后移除事实核查员并放宽内容审核）。毕竟美国自1996年来未通过重大联邦网络立法。
若司法反垄断路径受阻，克莱格推动的由美国主导、制定数据规则与人权标准的综合性协议，或将成为实现即时改进的唯一途径。
归根结底，治愈互联网痼疾难有单一解方。但三位作者的共识——增强用户控制权、保障数据隐私、强化硅谷责任——无疑值得世人共同追寻。
内森·史密斯，其作品见于《华盛顿邮报》《经济学人》《洛杉矶时报》等媒体。
深度聚焦
人工智能
与AI聊天机器人建立情感联结竟如此简单
当人类与聊天机器人的羁绊日益加深，对某些群体是安全保障，对另一些群体却暗藏危机。
谷歌首次披露AI指令能耗数据
这家科技巨头迄今最透明的能耗评估，为研究者提供了期待已久的内幕窥探。
心理治疗师秘密使用ChatGPT引发客户应激反应
某些治疗师在诊疗过程中使用AI，正在透支客户的信任与隐私。
为何人工智能的突破应该感谢鸽子
这种鸟类从未因智慧获得赞誉，但支撑尖端AI系统的强化学习技术，其思维模式更接近鸽子而非人类。
保持连接
获取《麻省理工科技评论》最新动态
发现特别优惠、热点资讯、活动预告与更多精彩内容。
英文来源：
Can we repair the internet?
Three new books propose remedies that run the gamut from government regulation to user responsibility.
From addictive algorithms to exploitative apps, data mining to misinformation, the internet today can be a hazardous place. Books by three influential figures—the intellect behind “net neutrality,” a former Meta executive, and the web’s own inventor—propose radical approaches to fixing it. But are these luminaries the right people for the job? Though each shows conviction, and even sometimes inventiveness, the solutions they present reveal blind spots.
In The Age of Extraction: How Tech Platforms Conquered the Economy and Threaten Our Future Prosperity, Tim Wu argues that a few platform companies have too much concentrated power and must be dismantled. Wu, a prominent Columbia professor who popularized the principle that a free internet requires all online traffic to be treated equally, believes that existing legal mechanisms, especially anti-monopoly laws, offer the best way to achieve this goal.
Pairing economic theory with recent digital history, Wu shows how platforms have shifted from giving to users to extracting from them. He argues that our failure to understand their power has only encouraged them to grow, displacing competitors along the way. And he contends that convenience is what platforms most often exploit to keep users entrapped. “The human desire to avoid unnecessary pain and inconvenience,” he writes, may be “the strongest force out there.”
He cites Google’s and Apple’s “ecosystems” as examples, showing how users can become dependent on such services as a result of their all-encompassing seamlessness. To Wu, this isn’t a bad thing in itself. The ease of using Amazon to stream entertainment, make online purchases, or help organize day-to-day life delivers obvious gains. But when powerhouse companies like Amazon, Apple, and Alphabet win the battle of convenience with so many users—and never let competitors get a foothold—the result is “industry dominance” that must now be reexamined.
The measures Wu advocates—and that appear the most practical, as they draw on existing legal frameworks and economic policies—are federal anti-monopoly laws, utility caps that limit how much companies can charge consumers for service, and “line of business” restrictions that prohibit companies from operating in certain industries.
Columbia University’s Tim Wu shows how platforms have shifted from giving to users to extracting from them. He argues that our failure to understand their power has only encouraged them to grow.
Anti-monopoly provisions and antitrust laws are effective weapons in our armory, Wu contends, pointing out that they have been successfully used against technology companies in the past. He cites two well-known cases. The first is the 1960s antitrust case brought by the US government against IBM, which helped create competition in the computer software market that enabled companies like Apple and Microsoft to emerge. The 1982 AT&T case that broke the telephone conglomerate up into several smaller companies is another instance. In each, the public benefited from the decoupling of hardware, software, and other services, leading to more competition and choice in a technology market.
But will past performance predict future results? It’s not yet clear whether these laws can be successful in the platform age. The 2025 antitrust case against Google—in which a judge ruled that the company did not have to divest itself of its Chrome browser as the US Justice Department had proposed—reveals the limits of pursuing tech breakups through the law. The 2001 antitrust case brought against Microsoft likewise failed to separate the company from its web browser and mostly kept the conglomerate intact. Wu noticeably doesn’t discuss the Microsoft case when arguing for antitrust action today.
Nick Clegg, until recently Meta’s president of global affairs and a former deputy prime minister of the UK, takes a position very different from Wu’s: that trying to break up the biggest tech companies is misguided and would degrade the experience of internet users. In How to Save the Internet: The Threat to Global Connection in the Age of AI and Political Conflict, Clegg acknowledges Big Tech’s monopoly over the web. But he believes punitive legal measures like antitrust laws are unproductive and can be avoided by means of regulation, such as rules for what content social media can and can’t publish. (It’s worth noting that Meta is facing its own antitrust case, involving whether it should have been allowed to acquire Instagram and WhatsApp.)
Clegg also believes Silicon Valley should take the initiative to reform itself. He argues that encouraging social media networks to “open up the books” and share their decision-making power with users is more likely to restore some equilibrium than contemplating legal action as a first resort.
But some may be skeptical of a former Meta exec and politician who worked closely with Mark Zuckerberg and still wasn’t able to usher in such changes to social media sites while working for one. What will only compound this skepticism is the selective history found in Clegg’s book, which briefly acknowledges some scandals (like the one surrounding Cambridge Analytica’s data harvesting from Facebook users in 2016) but refuses to discuss other pertinent ones. For example, Clegg laments the “fractured” nature of the global internet today but fails to acknowledge Facebook’s own role in this splintering.
Breaking up Big Tech through antitrust laws would hinder innovation, says Clegg, arguing that the idea “completely ignores the benefits users gain from large network effects.” Users stick with these outsize channels because they can find “most of what they’re looking for,” he writes, like friends and content on social media and cheap consumer goods on Amazon and eBay.
Wu might concede this point, but he would disagree with Clegg’s claims that maintaining the status quo is beneficial to users. “The traditional logic of antitrust law doesn’t work,” Clegg insists. Instead, he believes less sweeping regulation can help make Big Tech less dangerous while ensuring a better user experience.
Clegg has seen both sides of the regulatory coin: He worked in David Cameron’s government passing national laws for technology companies to follow and then moved to Meta to help the company navigate those types of nation-specific obligations. He bemoans the hassle and complexity Silicon Valley faces in trying to comply with differing rules across the globe, some set by “American federal agencies” and others by “Indian nationalists.”
But with the resources such companies command, surely they are more than equipped to cope? Given that Meta itself has previously meddled in access to the internet (such as in India, whose telecommunications regulator ultimately blocked its Free Basics internet service for violating net neutrality rules), this complaint seems suspect coming from Clegg. What should be the real priority, he argues, is not any new nation-specific laws but a global “treaty that protects the free flow of data between signatory countries.”
What the former Meta executive Nick Clegg advocates—unsurprisingly—is not a breakup of Big Tech but a push for it to become “radically transparent.”
Clegg believes that these nation-specific technology obligations—a recent one is Australia’s ban on social media for people under 16—usually reflect fallacies about the technology’s human impact, a subject that can be fraught with anxiety. Such laws have proved ineffective and tend to taint the public’s understanding of social networks, he says. There is some truth to his argument here, but reading a book in which a former Facebook executive dismisses techno-determinism—that is, the argument that technology makes people do or think certain things—may be cold comfort to those who have seen the harm technology can do.
In any case, Clegg’s defensiveness about social networks may not gain much favor from users themselves. He stresses the need for more personal responsibility, arguing that Meta doesn’t ever intend for users to stay on Facebook or Instagram endlessly: “How long you spend on the app in a single session is not nearly as important as getting you to come back over and over again.” Social media companies want to serve you content that is “meaningful to you,” he claims, not “simply to give you a momentary dopamine spike.” All this feels disingenuous at best.
What Clegg advocates—unsurprisingly—is not a breakup of Big Tech but a push for it to become “radically transparent,” whether on its own or, if necessary, with the help of federal legislators. He also wants platforms to bring users more into their governance processes (by using Facebook’s model of community forums to help improve their apps and products, for example). Finally, Clegg also wants Big Tech to give users more meaningful control of their data and how companies such as Meta can use it.
Here Clegg shares common ground with the inventor of the web, Tim Berners-Lee, whose own proposal for reform advances a technically specific vision for doing just that. In his memoir/manifesto This Is for Everyone: The Unfinished Story of the World Wide Web, Berners-Lee acknowledges that his initial vision—of a technology he hoped would remain open-source, collaborative, and completely decentralized—is a far cry from the web that we know today.
If there’s any surviving manifestation of his original project, he says, it’s Wikipedia, which remains “probably the best single example of what I wanted the web to be.” His best idea for moving power from Silicon Valley platforms into the hands of users is to give them more data control. He pushes for a universal data “pod” he helped develop, known as “Solid” (an abbreviation of “social linked data”).
The system—which was originally developed at MIT—would offer a central site where people could manage data ranging from credit card information to health records to social media comment history. “Rather than have all this stuff siloed off with different providers across the web, you’d be able to store your entire digital information trail in a single private repository,” Berners-Lee writes.
The Solid product may look like a kind of silver bullet in an age when data harvesting is familiar and data breaches are rampant. Placing greater control with users and enabling them to see “what data [i]s being generated about them” does sound like a tantalizing prospect.
But some people may have concerns about, for example, merging their confidential health records with data from personal devices (like heart rate info from a smart watch). No matter how much user control and decentralization Berners-Lee may promise, recent data scandals (such as cases in which period-tracking apps misused clients’ data) may be on people’s minds.
Berners-Lee believes that centralizing user data in a product like Solid could save people time and improve daily life on the internet. “An alien coming to Earth would think it was very strange that I had to tell my phone the same things again and again,” he complains about the experience of using different airline apps today.
With Solid, everything from vaccination records to credit card transactions could be kept within the digital vault and plugged into different apps. Berners-Lee believes that AI could also help people make more use of this data—for example, by linking meal plans to grocery bills. Still, if he’s optimistic on how AI and Solid could coordinate to improve users’ lives, he is vague on how to make sure that chatbots manage such personal data sensitively and safely.
Berners-Lee generally opposes regulation of the web (except in the case of teenagers and social media algorithms, where he sees a genuine need). He believes in internet users’ individual right to control their own data; he is confident that a product like Solid could “course-correct” the web from its current “exploitative” and extractive direction.
Of the three writers’ approaches to reform, it is Wu’s that has shown some effectiveness of late. Companies like Google have been forced to give competitors some advantage through data sharing, and they have now seen limits on how their systems can be used in new products and technologies. But in the current US political climate, will antitrust laws continue to be enforced against Big Tech?
Clegg may get his way on one issue: limiting new nation-specific laws. President Donald Trump has confirmed that he will use tariffs to penalize countries that ratify their own national laws targeting US tech companies. And given the posture of the Trump administration, it doesn’t seem likely that Big Tech will see more regulation in the US. Indeed, social networks have seemed emboldened (Meta, for example, removed fact-checkers and relaxed content moderation rules after Trump’s election win). In any case, the US hasn’t passed a major piece of federal internet legislation since 1996.
If using anti-monopoly laws through the courts isn’t possible, Clegg’s push for a US-led omnibus deal—setting consensual rules for data and acceptable standards of human rights—may be the only way to make some more immediate improvements.
In the end, there is not likely to be any single fix for what ails the internet today. But the ideas the three writers agree on—greater user control, more data privacy, and increased accountability from Silicon Valley—are surely the outcomes we should all fight for.
Nathan Smith is a writer whose work has appeared in the Washington Post, the Economist, and the Los Angeles Times.
Deep Dive
Artificial intelligence
It’s surprisingly easy to stumble into a relationship with an AI chatbot
We’re increasingly developing bonds with chatbots. While that’s safe for some, it’s dangerous for others.
In a first, Google has released data on how much energy an AI prompt uses
It’s the most transparent estimate yet from one of the big AI companies, and a long-awaited peek behind the curtain for researchers.
Therapists are secretly using ChatGPT. Clients are triggered.
Some therapists are using AI during therapy sessions. They’re risking their clients’ trust and privacy in the process.
Why we should thank pigeons for our AI breakthroughs
The bird has never gotten much credit for being intelligent. But the reinforcement learning powering the world’s most advanced AI systems is far more pigeon than human.
Stay connected
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.