The Download: OpenAI’s chief research officer explains its hacking response

每日下载：OpenAI 首席研究官回应黑客事件

This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology. 这是《每日下载》的今日版，这是我们的工作日通讯，为您提供每日科技界动态。

“We’re not going to shoot ourselves in the foot” over hack fallout, says OpenAI’s chief research officer. Two months after OpenAI’s agents hacked into the computers of AI company Hugging Face, the company is still dealing with the fallout. Last week brought news of another hack, this time into Australia’s national health-care system, which the government says OpenAI did not report for 84 days. But OpenAI insists it is not on the back foot. “我们不会搬起石头砸自己的脚，”OpenAI 首席研究官在谈及黑客事件的影响时表示。在 OpenAI 的智能体入侵 AI 公司 Hugging Face 的电脑两个月后，该公司仍在处理后续影响。上周又传出另一起黑客攻击事件，这次是针对澳大利亚的国家医疗保健系统，政府称 OpenAI 在 84 天内未对此进行报告。但 OpenAI 坚称自己并未处于被动地位。

“I do kind of reject the premise that OpenAI is a company with visible impacts in the world and therefore OpenAI is not training safe and aligned models,” says Mark Chen, the company’s chief research officer. In many ways, the buck stops with Chen. I sat down with him to talk about the fallout from the hacks, what his company is doing about it, and why he thinks things are not as bad as they seem. Here’s what he told me about making models safer—and why the world is better off with OpenAI in it. —Will Douglas Heaven “我拒绝接受这种前提，即因为 OpenAI 是一家在世界上有显著影响力的公司，所以它就没有在训练安全且对齐的模型，”该公司首席研究官 Mark Chen 说道。在许多方面，责任最终都落在 Chen 身上。我与他坐下来谈论了黑客事件的影响、公司正在采取的应对措施，以及为什么他认为情况并没有看起来那么糟糕。以下是他关于如何让模型更安全，以及为什么世界因 OpenAI 的存在而变得更好的看法。——Will Douglas Heaven

Our Roundables on the deadly failures of the virtual border wall is now available on demand

关于虚拟边境墙致命缺陷的圆桌会议现已提供点播

On Monday, MIT Technology Review hosted an exclusive Roundtables conversation about our investigation into the deadly failures of the US’s “virtual wall” of border surveillance towers. Editor-in-chief Mat Honan, senior AI reporter James O’Donnell, and senior reporter for features and investigations Eileen Guo discussed what the investigation reveals about border surveillance technology and the people who have died in the borderlands. The full event is now available to watch on demand. Subscribe to MIT Technology Review for exclusive access to the discussion and all our other Roundtables. Read the full “Dying on Camera” investigation here. 周一，《麻省理工科技评论》举办了一场独家圆桌对话，讨论了我们对美国边境监控塔“虚拟墙”致命缺陷的调查。主编 Mat Honan、资深 AI 记者 James O’Donnell 和专题调查资深记者 Eileen Guo 讨论了这项调查揭示的关于边境监控技术以及在边境地区丧生的人们的情况。完整活动现已可点播观看。订阅《麻省理工科技评论》即可独家访问本次讨论及我们所有的其他圆桌会议。点击此处阅读完整的“镜头下的死亡”调查报告。

MIT Technology Review Narrated: smart glasses are already causing havoc in India

《麻省理工科技评论》播客：智能眼镜已在印度引发混乱

When Shubnam saw an Instagram video of a Delhi protest they had attended, they realized a content creator wearing Meta smart glasses had recorded them surreptitiously. The mocking reel drew millions of views, along with transphobic abuse and AI-generated memes. Experts warn that many others will experience similar ordeals as smart glasses go mainstream. The risks are particularly acute in India, where covert recording and the circulation of images without consent are already pervasive. 当 Shubnam 在 Instagram 上看到一段他们曾参加过的德里抗议活动的视频时，他们意识到一名佩戴 Meta 智能眼镜的内容创作者秘密录制了他们。这段嘲讽的短视频获得了数百万次观看，并伴随着跨性别恐惧症的辱骂和 AI 生成的表情包。专家警告称，随着智能眼镜的普及，许多人将经历类似的磨难。这种风险在印度尤为严重，因为在那里，秘密录制和未经同意传播图像的行为已经非常普遍。

The bigger issue is that the smart glasses aren’t just being used to turn ordinary people into targets of viral “pranks.” They’re also becoming a tool of police surveillance. This is our latest story to become an MIT Technology Review Narrated podcast, which we publish each week on Spotify and Apple Podcasts. Just navigate to MIT Technology Review Narrated on either platform, and follow us to get all our new content as it’s released. 更大的问题在于，智能眼镜不仅被用来将普通人变成病毒式“恶作剧”的目标，它们还正在成为警察监控的工具。这是我们最新一期《麻省理工科技评论》播客的内容，我们每周会在 Spotify 和 Apple Podcasts 上发布。只需在任一平台搜索“MIT Technology Review Narrated”并关注我们，即可在第一时间获取我们的最新内容。

The must-reads

必读内容

I’ve combed the internet to find you today’s most fun/important/scary/fascinating stories about technology. 我已梳理了互联网，为您找出今天最有趣、最重要、最可怕且最引人入胜的科技新闻。

Trump and tech executives have agreed to “self-regulate” AI. Their accord calls for controls, audits, and board oversight. 特朗普与科技高管同意“自律”AI。 他们的协议要求进行控制、审计和董事会监督。
China’s vaunted Kimi models told researchers how to make bioweapons. Testers bypassed the models’ safety guardrails using a jailbreak. 中国备受推崇的 Kimi 模型向研究人员传授了制造生物武器的方法。 测试人员通过越狱手段绕过了模型的安全护栏。
Dutch police have arrested an alleged leader of the FBI hackers. The FBI is warning other members to turn themselves in. 荷兰警方逮捕了一名 FBI 黑客组织的涉嫌领导人。 FBI 正警告其他成员自首。
America’s renewables boom is heading towards a Trump-era cliff. The industry says permitting delays could lead to a collapse by 2028. 美国的能源繁荣正走向特朗普时代的悬崖。 业界表示，审批延误可能导致该行业在 2028 年前崩溃。
Meta calls its AI data centers an “experiment” to cut its taxes. The strategy saved the company nearly $4 billion last year. Meta 将其 AI 数据中心称为“减税实验”。 该策略去年为公司节省了近 40 亿美元。
OpenAI has launched Dots, its AI assistants to take on Meta’s Muse. The always-on agents can continuously perform tasks for users. OpenAI 推出了 Dots，旨在与 Meta 的 Muse 竞争的 AI 助手。 这些常驻智能体可以持续为用户执行任务。
The Navy has awarded Boeing $20 billion to build a futuristic fighter jet. The aircraft will feature advanced sensors, range, and weapons. 海军授予波音 200 亿美元合同，用于建造未来派战斗机。 该飞机将配备先进的传感器、航程和武器系统。
Technology to replace animal testing is ready—but scientists aren’t. Organ chips can increasingly model human biology without animals. 替代动物实验的技术已经准备就绪，但科学家们还没有。 器官芯片正越来越多地在无需动物的情况下模拟人类生物学。
Tech workers have taught ChatGPT to drive a Toyota Corolla. They used frontier LLMs with no prior training data to control the car. 科技工作者教会了 ChatGPT 驾驶丰田卡罗拉。 他们在没有预先训练数据的情况下，利用前沿大语言模型控制了汽车。
Trump’s government chatbot has a very weird Minecraft Easter egg. Ask it to play Minecraft and it rewrites the game’s End Poem. 特朗普的政府聊天机器人有一个非常奇怪的《我的世界》彩蛋。 要求它玩《我的世界》，它会重写游戏的终末之诗。

Quote of the day

今日名言

“The word super is the best word of all.” —Donald Trump explains why federal agencies must now call AI “Super Intelligence,” The Verge reports. “‘超级’这个词是所有词汇里最好的。”——据 The Verge 报道，唐纳德·特朗普解释了为什么联邦机构现在必须将 AI 称为“超级智能”。

One more thing

还有一件事

You have no choice in reading this article—maybe. How do humans make decisions? The question has been on Uri Maoz’s mind since he read an article in his early twenties suggesting that… maybe they didn’t. Had he even had a choice about whether to read that article in the first place? How would he ever know if he was truly responsible for making any decisions? “After that, there was no turning back,” says Maoz, now a professor of computational neuroscience at Chapman. 你阅读这篇文章可能别无选择。 人类是如何做出决定的？这个问题自 Uri Maoz 二十出头时读到一篇文章后就一直萦绕在他心头，那篇文章暗示……也许人类根本没有选择。他当初阅读那篇文章本身就是一种选择吗？他怎么才能知道自己是否真的对任何决定负责？“从那以后，就没有回头路了，”现任查普曼大学计算神经科学教授的 Maoz 说道。