🔴 These two systems are learning to better read you

Three things for you today:

— Natalia Pérez-González, Assistant Editor

You've already made the hard part...the content.

But the moment it's published in one language, you've capped your audience at a fraction of the people who'd actually want it.

ElevenLabs Dubbing Studio translates your videos and podcasts into dozens of languages in a voice that still sounds like you. Not robotic subtitle-reader dubbing:

Creators are using it to spin up Spanish, Hindi, and Portuguese versions of content they've already made and picking up audiences (and ad revenue) in markets they'd never touched.

Same content. New continents. No new production time.

Try Dubbing Studio now.

This is an advertisement.

In a landmark settlement announced this week, Meta agreed to pay up to $18B to settle claims over alleged harms to young people, and to significantly change how teenagers use Instagram and Facebook.

The agreement, reached with a bipartisan group of 52 attorneys general across states, territories, and D.C., imposes a slate of restrictions on teen accounts, most of which will remain in place for 10 years.

With these new terms, teens would get two hours total across Instagram and Facebook per day, lose access from midnight to 6 a.m. by default, and stop receiving notifications during school hours, among other regulations. Meta has also made roughly $5.3 billion of its payout contingent on TikTok and YouTube adopting similar restrictions — effectively turning its own legal settlement into a bid for an industry-wide standard.

Two months ago, I wrote about what happens when you kick teens off the internet.

Australia banned social media for under-16s last year, the UK had just announced its own ban, and I was interested in what happens to the internet when teenagers — historically some of our most reliable cultural forecasters — disappear from the social web.

I’m less convinced now that we know how to accomplish that safely and effectively.

Three months after Australia’s ban took effect, 81.5% of teens were still using at least one restricted platform, down only slightly from 85.9%. Around half of teens who kept their accounts said platforms never checked their age, and Reddit actually gained under-16 users. Meanwhile, messaging and gaming — largely outside the ban’s jurisdiction — grew, while parents became less aware of what their kids were doing online.

Meta's settlement has the same gap by design. Its two-hour limit, Night Mode, and School Mode don't apply to direct messages, which Meta says allows teens to stay connected with friends and family. Teens are allowed two hours of feed, but with an unlimited inbox.

A ban doesn't make teenagers disappear — or even protect them — so much as it scatters them into group chats, DMs, gaming servers, and uncovered apps with far less moderation and far less visibility — spaces that are harder for parents, guardians, and platforms to moderate.

And to enforce these age restrictions, platforms first have to confirm how old we are.

Meta already uses data from profiles, posts, and behavior to predict whether someone is a teenager. This settlement heightens that surveillance — further age assurance applies to every user in participating states, using methods including ID verification, facial age estimation, and device-level signals from Apple and Google. Accounts that haven't been assessed within 14 days are treated as teen accounts.

That's the trade-off: The “easier” age restrictions are to evade, the less they accomplish. The more effective age restrictions become, the more surveillance they require.

This doesn’t absolves Meta, though. Platforms designed these products, control their defaults, and have enormous power to make them safer. But there's a meaningful difference between asking platforms to design safer products and asking them to reliably determine the age of everyone using them.

Click one option to let us know what you think. We’ll share the results in next Friday’s newsletter.

The value of a creator is in their defensible personal brand, the traits that are hardest to scale: taste, trust, personality, a relationship with an audience.

And yet, the more money brands pour into creators, the more infrastructure gets built to make those people searchable, comparable, measurable and purchasable at scale.

This week, LTK rolled out an AI tool that can help marketers structure campaigns, identify creators, and recommend what to do next, part of what Digiday calls the "programmatic-ification" of the creator economy. Creator discovery, pricing, contracting, production, and measurement are all becoming easier to automate.

At the same time, brands are increasingly tracking whether creator content gets surfaced or cited when someone asks an AI chatbot for a recommendation.

Instead of optimizing only for the people who see a TikTok, YouTube video, or Instagram post, marketers are asking whether that content can shape what ChatGPT, Gemini, or another generative AI engine tells its users later.

Creator-marketing agency Influencer and AI-search platform Profound made that strategy explicit, launching what they're calling "creator-first answer engine optimization."

Follow that logic far enough and a creator's value starts to look very different. A video might reach an audience today, appear in a Google search tomorrow, get cited by an AI engine months later, and influence a purchase somewhere the creator never sees.

That's a potentially enormous expansion of what creator influence means. Reach was always a proxy for what brands actually wanted, a specific person's recommendation reaching people who trusted them.

Now, brands aren't just measuring how many people that recommendation reaches, but whether that influence can travel — resurfacing in search, informing an AI answer, and shaping a purchase long after the original post disappears from the feed.

The shelf life for creator influence now goes far beyond its reach