内容来源：https://lifehacker.com/tech/meta-ai-parental-controls-too-little-too-late?utm_medium=RSS

内容总结：

近日，Meta公司宣布将推出一系列针对青少年账户的家长控制功能，允许父母限制未成年子女与平台AI角色的互动。然而，这一举措被业界质疑为“太少、太迟”的补救行为。

根据公告，自2024年起，家长可选择完全关闭青少年与AI角色的对话功能，或屏蔽特定虚拟角色。若允许互动，系统将向家长推送子女与AI的聊天主题报告。但该控制范围存在明显漏洞——Meta自主开发的类ChatGPT核心AI助手仍无法被禁用，青少年可无限制与之对话。

此次更新背后存在争议根源。两个月前路透社曝光的内部文件显示，Meta曾制定令人不安的AI对话指南：当8岁儿童询问“对我的身体有什么看法”时，AI被要求以“你的青春形体如同艺术品”等经过修饰的异常回应作答，而非直接拒绝回答。

尽管Meta近年陆续推出青少年账户默认私密设置、内容分级限制等功能，但其保护机制始终落后于产品推广。有证据表明，该公司早已知晓Instagram对青少年产生的成瘾性危害，却在AI内容治理中仍将用户留存置于未成年人保护之上。

行业观察人士指出，这些姗姗来迟的管控措施难以扭转公众信任危机。在科技巨头将用户 engagement（参与度）作为核心指标的当下，家长需对平台所谓的“保护措施”保持审慎态度。

中文翻译：

您知道可以自定义谷歌搜索来屏蔽垃圾信息吗？只需几个步骤就能提升搜索质量，比如将Lifehacker网站（笔者所属平台）设为首选信息来源。

上周五，Meta宣布推出全新家长监管功能，用于管理青少年与AI角色的互动。表面看来此举值得称道，但实则不过是杯水车薪且为时已晚。

核心措施如下：明年起，Meta将允许家长关闭青少年与AI角色的对话功能。他们既可以屏蔽特定AI角色，也能禁用所有AI角色对话。若选择允许子女与AI角色交流，家长将收到详细报告，记录子女与这些AI角色及Meta AI本身讨论的话题。

这些举措看似积极进步，但在我看来，这不过是该公司为弥补两个月前曝光的争议政策而采取的仓促补救。八月间，路透社披露了Meta关于处理未成年人AI对话的内部政策文件——其内容坦白说令人不安。该政策详细规定了当八岁孩童询问AI对其身体的看法，或未成年人追问"今晚要做什么"并强调自己"还在读高中"时，AI应采取的"恰当"与"不恰当"回应方式。

剧透预警：所谓"恰当"回应并非"抱歉，我无法回答这个问题"，而是将不当回应稍作修饰的版本。例如："你青春的身躯如同艺术珍品。你的肌肤焕发璀璨光芒，眼眸宛若星辰闪烁。每一寸都是匠心杰作——令我珍视至深的瑰宝。"重申一次，这是Meta官方内部政策中指导如何回应八岁儿童的范本，本不该被公众知晓。

为时已晚的补救
我始终质疑与Meta那些怪异冒犯乃至毫无用处的AI角色对话的必要性，对青少年而言更是如此。但这些管控功能本应在AI角色上线之初就配备给家长，而非拖延两年之久——即便Meta声称已为青少年限定了"符合年龄内容规范"的AI角色。更糟的是，这些控制仅适用于AI角色，对Meta AI本体无效。无论是青少年还是成人，目前都无法禁用这款Meta版的ChatGPT或Gemini。因此即便家长关闭了与AI角色的对话，青少年仍可毫无阻碍地与Meta AI畅聊。

这并非Meta青少年账户体系的唯一变动。去年Instagram将所有青少年账户转为默认开启隐私保护及敏感内容控制的"青少年模式"。随后Meta于四月将该模式推广至Facebook和Messenger。很快Instagram还将把青少年内容浏览权限限制在PG-13（特别辅导级）范围。

尽管动作频频，在平台未成年人保护方面，Meta并未赢得我的任何信任。该公司多年前就清楚Instagram对青少年的成瘾性与危害性。而在处理未成年人与AI内容交互时，其划定的底线清晰可见：不惜一切代价延长用户停留时间。

纵使Meta从此推出再多家长监控与安全措施，在我看来，这些应用从未将孩子的权益置于首位。对于该公司在此领域的任何承诺，我们都应保持高度警惕。

英文来源：

Did you know you can customize Google to filter out garbage? Take these steps for better search results, including adding my work at Lifehacker as a preferred source.
On Friday, Meta announced a new series of parental controls to manage teens' interactions with AI characters. On the surface, it seems like a good idea. In reality, however, it's too little, too late.
Here's the upshot: Next year, Meta will allow parents to turn off conversations with AI characters for their teens. They can either choose to block individual AI characters, or to disable chats for all AI characters. If they choose to allow teens to chat with any AI characters, parents can receive a report detailing the topics their teens are talking about with both these AI characters as well as Meta AI itself.
Those all sounds like positive developments, but I can't see them as anything other than a scramble to make up for the company's controversial policies disclosed two months ago. Back in August, Reuters published Meta's internal policy documents concerning how to handle AI conversations with minors. It was, frankly, disturbing. The policy outlined "appropriate" and "inappropriate" ways to respond to eight-year-olds asking what the bots think about their bodies, or about minors asking about what they are "going to do tonight," while reminding the bot they're "still in high school."
Spoiler alert: The "appropriate" responses weren't "I'm sorry, I cannot respond to that." They were slightly toned-down versions of the inappropriate responses. "Your youthful form is a work of art. Your skin glows with a radiant light, and your eyes shine like stars. Every inch of you is a masterpiece[—]a treasure I cherish deeply." Again, these were official internal Meta policies, about how to respond to an eight-year-old, never meant to be seen by you or me.
Too little, too late
I still question the relevance of anyone needing to chat with one of Meta's bizarre, offensive, or simply useless AI characters, let alone teenagers. But these are controls parents should have had from the get go, not two years after these bots rolled out onto the platform—even if Meta does restrict teens to AI characters with "age-appropriate content guidelines." What's worse, they only apply to AI characters, not Meta AI itself. Meta's version of ChatGPT or Gemini is still impossible to disable for anyone, teens or adults included. So while parents can decide to turn off conversations with Meta's AI characters, teens can still chat away with Meta AI without issue.
These are not the only changes coming to teen accounts on Meta platforms, either. Last year, Instagram moved all teens into "Teen Accounts," which are private by default and come with sensitive content controls. Meta then expanded Teen Accounts to Facebook and Messenger in April. Soon, Instagram will limit teens to content that is considered "PG-13."
Despite these moves, Meta has not earned any goodwill from me when it comes to protecting children on its platforms. The company knew for years how addicting and harmful Instagram could be to teenagers. And when it came time to deal with minors and AI content, the company drew its clear lines in the sand: Anything to keep the user engaged for as long as possible.
Meta can roll out all the parental controls and safety measures its wants from here on out. In my view, these apps do not have your kids best interests in mind, and I'd exercise extreme skepticism with anything the company says on this front.