Why We Shouldn’t Regulate Kids on Social Media (and Why We Should)

Hello world,

Thank you to everyone who attended the meetup! It was a blast. Subscribers will always be first to hear about events and get togethers. Looking forward to hosting more in the future.

Last week a bunch of court rulings and new laws all landed basically within 72 hours of each other, all aimed at social media’s relationship with kids. You might be surprised to hear that not everyone wants more regulation when it comes to kids and their online lives. I’ve seen a lot of debate coming from the two sides, with some people seeing regulation as long overdue, and others saying this isn’t addressing the problem at all, and maybe the situation is a discipline and culture problem being used to justify government overreach that will make Big Tech even more powerful.

Let me know what you think.

Here are some recent headlines covering kids and social media so you get the idea:

• March 24: A New Mexico jury awarded $375M after finding Meta willfully violated consumer protection laws. The state AG built the case using a fake 13-year-old profile and documented thousands of violations. First successful state lawsuit against Meta over child safety. (CNBC)

• March 25: A Los Angeles jury found Meta and YouTube negligent for failing to warn users about addictive platform design, awarding $6M in damages. First jury verdict ever holding tech companies responsible for social media addiction, with ~2K similar cases pending. (NPR)

• March 25: The UK government launched a six-week pilot testing social media restrictions on 300 teenagers (full ban, one-hour cap, 9pm-7am curfew, and a control group) to measure impacts on schoolwork, sleep, and family life. (CNBC)

• March 26: The European Commission opened a formal DSA investigation into Snapchat over inadequate age verification, exposure to grooming, unsafe default settings for minors, and the sale of banned products. The Commission called Snap’s system of asking users to confirm they’re 13 “insufficient.” Fines could reach 6% of annual revenue. (Reuters)

The case for regulation:

It’s pretty obvious to know why people would want regulation. The platforms knew what was going on when it came to kids not feeling the best after spending hours on social media. I can remember finding out that Instagram’s own internal research found that its product made body image issues worse for one third of teenage girls like it was yesterday. These companies were aware of the harm, it didn’t get fixed, and now the courts and regulators above are the result.

Derek Thompson came out with a great read related to this topic, focusing more on general use of smartphones, but I think it’s still relevant.

• Researchers “found that randomly removing internet access from smartphones produced a range of benefits, including improved mental health, subjective well-being, and the ability to sustain attention. More than 90 percent of the nearly 500 participants experienced at least one benefit.” The biggest reason? Participants spent more time “socializing in person, exercising, and being in nature.” (Thompson, citing 2025 randomized trial)

• A separate study that paid people to deactivate Facebook found “those who logged off were happier, less anxious, and less politically polarized.” They also “spent more time with friends and family.” (Thompson, citing Allcott et al. 2020)

• Thompson’s favorite framing, borrowed from Jonathan Haidt: “phones replace play-based adolescence with phone-based adolescence.” As Thompson puts it, “the most important thing about phones isn’t what’s on the screen, but rather everything that’s off the screen when you’re lost gazing into your pocket device.” (Thompson)

• He thinks both sides “should get comfortable thinking about social media as ‘attention alcohol’: fun if moderately problematic for most, and very dangerous for some.” (Thompson)

We regulate actual alcohol with age restrictions, advertising rules, and warning labels. Maybe we could start seeing a similar treatment here.

The case against it:

This is where things get spicier.

Ben Sperry wrote about the concept of “liability-driven curation,” which Adam Thierer flagged on X. The argument: all of these lawsuits and investigations are creating a new reality for platforms. “Moderate too little, and they risk consumer-protection suits. Build engaging features, and they invite claims of addiction. Resist government pressure, and they may trigger investigations or regulatory backlash.” The result? “When the cost of hosting speech rises high enough, platforms host less of it.”

Going back to Thompson, he suggests smartphones are “an active ingredient that’s interacting with other phenomena that are distinctly western or American, in order to create berserk local effects.” If that’s true, regulating the phone doesn’t fix the culture.

He also notes that even though we all have phones, we are only seeing the negative outcomes on our society in certain areas, it’s not a global thing. “Happiness among young people has plummeted most severely in Western developed countries that speak English, such as the United States, Great Britain, Canada, Australia, and New Zealand.” Meanwhile, “happiness at every age has risen sharply in Central and Eastern Europe” and in East Asia, happiness is increasing “at every age.” (Thompson, citing World Happiness Report) When it comes to mental health crisis and attention spans, Thompson found that “Emergency-room visits for suicide attempts among young women soared across the Anglosphere” but “the suicide rate among people ages 15 to 19 fell in most European countries.” ADHD diagnoses in the U.S. are “rising at roughly twice the rate of European countries.”

Brian Castrucci made a similar case in Forbes back in February. He argues bans don’t teach kids anything. “Skills like regulating attention, navigating social pressure, interpreting online content and managing emotions don’t suddenly appear upon one’s 16th birthday.” He compares it to food: tightly controlling what kids eat at home doesn’t build healthy habits, it just delays the problem until college. Castrucci also points out that much of the research linking social media to poor mental health is correlational, not causal. “Distress may drive use just as much as use drives distress.” His position: “Bans may feel like protection, but education is what actually protects.”

So where does this leave us?

We’re in the middle of a massive, messy, high-stakes experiment where governments are legislating, courts are ruling, and researchers are still arguing about the data, all at the same time. Whether that leads to a safer internet for kids or a more controlled internet for everyone... we’re about to find out.

TECH NEWS

More Kids Online

• The European Commission released preliminary DSA findings accusing Pornhub, Stripchat, XNXX, and XVideos of failing to protect minors from pornographic content after a 10-month investigation. The Commission found the platforms relied on a single click to confirm users are over 18 and, when they did identify risks in their own reports, focused on reputational damage to themselves rather than harm to children. The sites can respond before a final ruling; penalties could reach 6% of global annual revenue. (Reuters, European Commission, Politico, Gamereactor UK)

• Washington Gov. Bob Ferguson signed two AI bills into law. The first (HB 1170) requires large AI companies to embed watermarks or metadata in AI-generated content so it can be traced back as machine-made. The second creates rules for “companion chatbots” like ChatGPT and Claude, requiring them to disclose they are not human at the start of every conversation and every 3 hours (every hour for minors). The chatbot law also bans sexually explicit conversations with minors, prohibits “manipulative engagement techniques,” and requires companies to flag concerning conversations and connect users with mental health services. (OPB)

Chips

• Arm released its first in-house chip in 35 years: the AGI CPU, a 136-core data center processor designed for AI inference. Meta co-developed the chip and is the first customer, with OpenAI, Cloudflare, and SAP also signed on. Arm’s stock jumped 16% after the company projected $15B in revenue from the chip by 2031. The move shifts Arm from licensing chip designs to directly competing with the companies it has traditionally supplied. (TechCrunch, CNBC)

• Elon Musk announced Terafab, a $20-25B semiconductor plant in Austin, Texas, jointly backed by Tesla, SpaceX, and xAI. The facility would manufacture custom chips for EVs, Optimus robots, and AI computing. Musk framed the project as a step toward U.S. chip independence from TSMC. (Futurism)

Money Moves

• Harvey, the AI legal startup, raised $200M at an $11B valuation, up from $8B three months ago, bringing its total funding past $1B. The company builds AI tools that handle legal briefs, due diligence, and other work traditionally done by junior associates. Major law firms are among its paying customers. (Bloomberg)

• A group of former Anthropic, xAI, and Google researchers founded a startup called humans& and raised $480M at a $4.48B valuation, a record for a seed round. Co-founders include Georges Harik, Google’s seventh employee. The company’s stated mission is building AI focused on “human collaboration and connection,” positioning itself as an alternative to the capability-focused approach at frontier AI labs. (Tech Startups)

• Amazon made two robotics acquisitions in one week: Fauna Robotics, which makes a 3.5-foot-tall, 50-pound bipedal humanoid called “Sprout” priced at $50K, and Rivr, a Swiss startup building stair-climbing delivery robots that can carry 30kg. Fauna’s ~50 employees will join Amazon’s Personal Robotics Group; Sprout will remain available to outside researchers. (TechCrunch)

  

  Fauna

Scroll Stopping

• OpenAI indefinitely shelved plans for an adult chatbot, reportedly called “Citron mode,” after pushback from staff and investors over risks including unhealthy emotional dependence and minors’ exposure to explicit content. At least one senior employee resigned over the project. The company also faced technical hurdles: models trained to avoid sexual content proved difficult to retrain, and its age verification system has an error rate above 10%. The pause is part of a broader strategic shift ahead of a potential IPO later this year. (Financial Times)

• Apple CEO Tim Cook said on Good Morning America, “I don’t want people looking at their smartphone more than they’re looking in someone’s eyes,” urging users to reduce screen time. The comment drew attention given Apple’s position as the world’s largest smartphone maker. (Daily Dot)

• About 1 in 8 U.S. adults now takes a GLP-1 drug like Ozempic or Zepbound. Roughly 70% of users report snacking less; fast-food dinner traffic has dropped 6% among regular users. Restaurants and food companies are responding with protein- and fiber-heavy reformulations to appeal to a customer base that is consuming less overall. (CNBC)

Workplace

• Meta cut roughly 700 jobs across Reality Labs, recruiting, sales, and parts of Facebook. Reality Labs, the division responsible for Quest headsets and metaverse development, has now gone through multiple rounds of layoffs while the unit continues to lose billions per quarter. The cuts came the same week Zuckerberg was named to the President’s science advisory council. (CNBC, TechCrunch, Bloomberg)

• Meta began offering top executives stock options tied to future price milestones, a compensation structure it hasn’t used since its 2012 IPO. The move is widely seen as a retention play amid intense competition for AI talent. The timing coincides with the company’s latest round of layoffs affecting 700 employees. (Bloomberg, CNBC)

• Block has quietly begun rehiring some of the employees Jack Dorsey laid off last month when the company cut 40% of its workforce, citing AI as the reason. At least one engineer was told he had been let go “by mistake.” The rehires involve a small number of former employees. (TheStreet)

• AI companies are driving Manhattan’s commercial real estate recovery, adding roughly 1M sqft of office space in 2025, a 152% jump. Anthropic, OpenAI, and Palantir are among the tenants, paying an average of $88/sqft compared to the $78 citywide average. (Bloomberg)

The White House

• The White House appointed Mark Zuckerberg, Jensen Huang, and Larry Ellison to the President’s Council of Advisors on Science and Technology, co-chaired by AI czar David Sacks and senior tech adviser Michael Kratsios. Other appointees include Marc Andreessen, Sergey Brin, and Lisa Su. The appointments place leaders of major tech companies in a formal advisory role to the administration on technology policy. (Bloomberg, Fortune, White House)

• First Lady Melania Trump was escorted into the White House East Room by Figure 03, a humanoid robot built by Figure AI, during the second day of her international tech summit. The robot addressed spouses of leaders from 45 nations in 11 languages. The summit, “Fostering the Future Together,” included representatives from 28 tech companies discussing AI and education for children. (CNN)