Kids’ Content Is Deeply Broken. A New Platform Wants to Fix It

本文为原文前 6,000 字符的节选翻译,完整内容请查看原文。

Kids’ Content Is Deeply Broken. A New Platform Wants to Fix It

If you have young kids or if you know somebody who does, then you know about YouTube’s AI slop problem. Videos targeted at children on the platform are often filled with bizarre and hyperstimulating content that’s AI generated. Not all of it is age-appropriate. These videos autoplay on endless, algorithmically curated streams punctuated by ads.

如果你有年幼的孩子,或者认识有孩子的人,那么你一定了解 YouTube 上充斥着 AI 生成的劣质内容(AI slop)这一问题。该平台上针对儿童的视频往往充斥着由人工智能生成的怪异且过度刺激的内容,其中并非所有内容都适合儿童观看。这些视频会在算法策划的无尽流中自动播放,并穿插着广告。

Maka, a new streaming app for kids aged 0 to 6, is positioning itself as an antidote to that concern. The platform has zero ads and doesn’t allow autoplay or endless scrolling, and its algorithms intentionally do not maximize watch time. The company’s cofounder and chief product officer, Tanyella Leta, tells WIRED the platform is purposely abandoning the mechanisms that “turn children’s media into an engagement machine.”

Maka 是一款面向 0 到 6 岁儿童的新型流媒体应用,它将自己定位为解决这一问题的良药。该平台没有任何广告,不允许自动播放或无尽滚动,其算法也刻意不追求最大化观看时长。该公司联合创始人兼首席产品官 Tanyella Leta 对《连线》(WIRED)表示,该平台正有意摒弃那些“将儿童媒体变成参与度机器”的机制。

Maka’s CEO and cofounder Isabel Sheinman adds that a lot of the streamers were built to serve adults, not children. “Kids were really the afterthought,” she says. Sheinman wants to set a much higher bar for what quality looks like in children’s media, which she says has been warped by harmful incentives. “Maka was built with child development science and children’s behavior first and foremost.”

Maka 的首席执行官兼联合创始人 Isabel Sheinman 补充说,许多流媒体平台最初是为成年人而非儿童服务的。“孩子们实际上是事后才被考虑到的,”她说。Sheinman 希望为儿童媒体的质量设定更高的标准,她认为目前的标准已被有害的激励机制所扭曲。“Maka 的构建始终将儿童发展科学和儿童行为放在首位。”

Launching today, Maka’s programming slate is a mix of licensed content from global studios and independent creators. It features 25 shows in its curated catalog, including Franklin, The Wiggles, Pocoyo, Vida the Vet, and Sesame Street. The platform runs on a subscription service, with a $15 monthly plan and a $107 annual plan.

Maka 于今日上线,其节目阵容结合了来自全球工作室和独立创作者的授权内容。其精选目录中包含 25 部剧集,包括《富兰克林》(Franklin)、《威格尔斯》(The Wiggles)、《小P优优》(Pocoyo)、《兽医维达》(Vida the Vet)和《芝麻街》(Sesame Street)。该平台采用订阅制,月费为 15 美元,年费为 107 美元。

It was during the Covid pandemic that Sheinman and Leta became obsessed with thinking about how to redesign kids’ media. The solution? Every episode streamed on Maka is evaluated by child development experts through Imprint, a framework created with researchers at the Yale Child Study Center. The evaluation process, which examines everything from scene cuts to how fast-paced content is, maps child development across seven core domains. It looks for things like emotional regulation, kindness, how child-caregiver relationships are modeled in a story, and how a story might support critical thinking or language development.

正是在新冠疫情期间,Sheinman 和 Leta 开始痴迷于思考如何重新设计儿童媒体。解决方案是什么呢?Maka 上播放的每一集内容都由儿童发展专家通过“Imprint”框架进行评估,该框架是与耶鲁大学儿童研究中心的专家共同创建的。评估过程涵盖了从场景剪辑到内容节奏快慢等各个方面,并对照儿童发展的七个核心领域进行映射。它关注诸如情绪调节、善良品质、故事中如何塑造儿童与照料者的关系,以及故事如何支持批判性思维或语言发展等要素。

AI-generated content has only added a new wrinkle to kids’ entertainment: videos are incoherent, have no plot, or contain outright misinformation but are packaged to look like learning. Some of it, as WIRED has previously reported, is disturbingly sexualized, featuring gore and other fetishes.

AI 生成的内容为儿童娱乐增添了新的麻烦:视频逻辑不通、没有情节,甚至包含明显的错误信息,却被包装成学习内容。正如《连线》此前报道的那样,其中一些内容令人不安地色情化,甚至包含血腥和其他恋物癖元素。

Michael Robb, head of research at the media literacy nonprofit Common Sense Media, says that on a platform like YouTube, slop videos that have the veneer of education sit side by side with carefully designed educational shows. Robb believes YouTube and other kids’ streaming platforms have not done enough to distinguish educational content from slop, which has only made filtering content harder for parents. “Platforms have taken steps to reduce kids’ exposure to age-inappropriate content, and that matters. But ‘not inappropriate’ is a very low bar. There’s a big gulf between that and media created intentionally with kids in mind.”

媒体素养非营利组织 Common Sense Media 的研究主管 Michael Robb 表示,在 YouTube 这样的平台上,披着教育外衣的劣质视频与精心设计的教育节目并存。Robb 认为,YouTube 和其他儿童流媒体平台在区分教育内容和劣质内容方面做得还不够,这只会让家长过滤内容变得更加困难。“平台已经采取措施减少儿童接触不适龄内容,这很重要。但‘不适龄’只是一个很低的门槛。这与专门为儿童设计的媒体之间存在巨大的鸿沟。”

A 2020 study by Common Sense Media on YouTube videos watched by young kids found 75 percent had “weak or no educational value.” Kids also frequently ended up watching ads and content intended for older audiences.

Common Sense Media 2020 年针对幼儿观看的 YouTube 视频进行的一项研究发现,75% 的视频“教育价值微弱或毫无价值”。孩子们还经常最终观看到广告以及面向成年观众的内容。

According to its policy for YouTube Kids, which lets parents approve specific content themselves, YouTube excludes videos that are “deceptive, sensational, or clickbait” but acknowledges that not every video has been manually reviewed. YouTube also makes clear that automated filtering cannot guarantee every video will be appropriate. “There’s a chance that your child will find content that you don’t want them to watch,” the company notes in its support docs.

根据 YouTube Kids 的政策(该政策允许家长自行批准特定内容),YouTube 会排除“欺骗性、耸人听闻或点击诱饵”的视频,但也承认并非每个视频都经过人工审核。YouTube 还明确表示,自动过滤无法保证每个视频都合适。该公司在支持文档中指出:“您的孩子仍有可能找到您不希望他们观看的内容。”

Earlier this year, in an effort to encourage better digital literacy among kids, the American Academy of Pediatrics updated its guidance beyond screen time recommendations, urging children to avoid “AI-generated content,” “AI chatbots,” “short-form videos,” and “highly sensationalized content or clickbait.” The guide prioritized quality content and coviewing with parents rather than passive watching.

今年早些时候,为了鼓励儿童提高数字素养,美国儿科学会更新了指南,超越了单纯的屏幕时间建议,敦促儿童避免接触“AI 生成内容”、“AI 聊天机器人”、“短视频”以及“高度耸人听闻的内容或点击诱饵”。该指南优先推荐高质量内容和家长陪同观看,而非被动观看。

Tiffany Munzer, a developmental-behavioral pediatrician and digital media researcher at the University of Michigan, was the lead author on Academy of Pediatrics’s “Digital Ecosystems, Children, and Adolescents” policy statement, and tells WIRED in an email that the proliferation of low-quality content has led kids to “miss out on important, slower learning opportunities that promote critical thinking, generate new ideas, and practice stretching their attention spans.”

密歇根大学的发育行为儿科医生兼数字媒体研究员 Tiffany Munzer 是美国儿科学会“数字生态系统、儿童与青少年”政策声明的主要作者。她在给《连线》的一封电子邮件中表示,低质量内容的泛滥导致孩子们“错失了重要的、节奏较慢的学习机会,而这些机会本可以促进批判性思维、激发新想法并练习延长注意力跨度。”

In April, child advocacy group Fairplay sent a letter, signed by more than 200 organizations and individual experts, asking YouTube to address the spread of AI videos on its platforms, saying AI slop “harms children’s development by distorting their sense of reality, overwhelming their learning processes and hijacking their attention.”

今年 4 月,儿童权益倡导组织 Fairplay 发出一封由 200 多个组织和个人专家签署的信函,要求 YouTube 解决其平台上 AI 视频泛滥的问题,称 AI 劣质内容“通过扭曲儿童的现实感、压垮他们的学习过程并劫持他们的注意力,从而损害了儿童的发展。”

A YouTube spokesperson tells WIRED that on YouTube Kids, which is meant for kids 12 and under, AI content is limited to a small set of high-quality channels, and parents have the option to block channels entirely. The platform has stricter standards for content made for younger viewers, the company says.

YouTube 发言人告诉《连线》,在面向 12 岁及以下儿童的 YouTube Kids 上,AI 内容仅限于少数高质量频道,且家长可以选择完全屏蔽某些频道。该公司表示,平台对面向低龄观众的内容有更严格的标准。

For Maka’s business model to succeed, its shows will need to entertain and educate kids while also establishing confidence in parents. “The value that parents want is relief, knowing that what their child is seeing has been vetted by child development experts and everything on Maka they can trust is good for their kid,” Sheiman says.

为了使 Maka 的商业模式取得成功,其节目不仅需要娱乐和教育儿童,还需要建立家长的信任。“家长们想要的价值是安心,即知道他们的孩子所看的内容已经过儿童发展专家的审核,并且他们可以相信 Maka 上的所有内容对孩子都有益,”Sheinman 说。

Yalda Uhls, a former film executive at Sony and MGM who is now a researcher at UCLA focusing on how media affects the social learning of young children, worries that as AI revolutionizes kids’ streaming content, people who study child development are being left out of the process. “When I was a studio executive, every story meant for children passed through layers of accountable adults,” she says.

Yalda Uhls 曾是索尼和米高梅的电影高管,现在是加州大学洛杉矶分校的研究员,专注于媒体如何影响幼儿的社会化学习。她担心,随着 AI 彻底改变儿童流媒体内容,研究儿童发展的人却被排除在这一过程之外。“当我还是电影公司高管时,每一个面向儿童的故事都要经过多层负责任的成年人的审核,”她说。