PART ONE: Meta & YouTube — The Dam Is Breaking
What Actually Happened (Two Verdicts in Two Days)
Verdict 1 — New Mexico, March 24: A Santa Fe jury
found Meta willfully violated New Mexico's consumer protection laws and ordered
the social media giant to pay $375 million in damages. The case centered on
child sexual exploitation — investigators created accounts on Facebook and
Instagram posing as users younger than 14, and those accounts received sexually
explicit material and were contacted by adults seeking similar content.
Critically, internal messages from Meta employees discussed how CEO Mark
Zuckerberg's 2019 announcement to make Facebook Messenger end-to-end encrypted
by default would impact the ability to disclose to law enforcement some 7.5
million child sexual abuse material reports.
Verdict 2 — Los Angeles, March 25: A jury found Meta
and YouTube negligent for designing apps that harmed kids and awarded $3
million in compensatory damages, plus an additional $3 million in punitive
damages — bringing the total to $6 million. Meta would pay $4.2 million and
YouTube $1.8 million.
Why These Are Bigger Than the Dollar Amounts Suggest
The individual payouts are almost irrelevant to Meta's
balance sheet — the fine is a tiny fraction of Meta's $201 billion revenue in
2025, and Meta's stock was actually up 5% after the New Mexico verdict.
What matters is the legal theory that just got validated.
The verdict validated the plaintiff's lawyers' approach of shifting the legal
target — instead of focusing on the content people see on social media, the
case put the spotlight on how social media services were designed. Meta's apps
were deliberately built to be addictive, and executives knew this and failed to
protect their youngest users.
This is the unlock. Once you establish design defect
as the liability hook, you bypass Section 230 — the long-standing shield that
protected platforms from being sued over user content.
What Comes Next: The Litigation Tsunami
Kaley's case was the first of more than 1,500 similar cases
against the social media companies to go to trial. Repeated losses could put
the tech giants on the hook for up to billions of dollars and force them to
change their platforms. The companies are also set to stand trial later this
year in the first of hundreds of additional lawsuits brought by school
districts and state attorneys general from around the country.
The landmark verdict may influence the outcome of 2,000
other pending lawsuits.
More than 40 state attorneys general have filed lawsuits
against Meta, claiming it's contributing to a mental health crisis among young
people by deliberately designing Instagram and Facebook features that are
addictive.
Here's the Charlie Munger inversion — what's the
worst realistic outcome for Meta and Google?
The Big Tobacco playbook. Some watching the Los
Angeles and other lawsuits move forward have anticipated a "Big Tobacco
moment" — a reference to the 1990s lawsuits against tobacco companies that
proved they were aware of the addictive nature of nicotine and the health dangers
of smoking, and led to massive damages paid. The tobacco master settlement
ended up costing the industry $246 billion spread over 25 years. That's
the ceiling scenario here.
The specific escalation path looks like this:
- More
bellwether trials, 2026 — Each win by plaintiffs increases settlement
pressure on the remaining 1,500+ individual cases. Lawyers will now flood
the zone with filings.
- School
district trials — These are potentially more damaging because
districts can aggregate harm across thousands of students and have deep
documentary evidence of mental health crises.
- State
AG trials — New Mexico's second trial phase begins May 4, when AG
Torrez will bring a public nuisance claim before a judge and seek
injunctive relief including real age verification, algorithm changes, an
independent monitor, and fundamental changes to how Meta does business in
the state. If New Mexico wins that, every other state AG gets a
template.
- Federal
legislation — Two successive jury verdicts in two days, with internal
documents showing executives knew, creates enormous political pressure on
Congress to finally pass platform liability reform.
- International
exposure — The EU, UK, and Australia already have tougher child safety
frameworks. These verdicts will accelerate enforcement actions abroad.
The real cost isn't even fines — it's forced product
redesign. Removing infinite scroll, algorithmic recommendation for minors,
auto-play, and engagement-maximizing features would be existential changes to
the core business model. Instagram without the recommendation engine is a
fundamentally less addictive, less profitable product.
π€ PART TWO: Melania's
Robot and the Zuck AI Education Frontier
What Actually Happened
Melania Trump walked side by side with a humanoid robot
called "Figure 03" — built by Sunnyvale-based startup Figure AI —
down a red carpet at the White House for her "Fostering the Future
Together" global coalition summit, which brought together first spouses
from around the world to discuss empowering children through educational
technology including AI.
The first lady invited attendees to imagine a "humanoid
educator named Plato" who could teach classical studies, saying the use of
robots would give children more time to be with friends, play sports and
develop extracurricular interests — producing "a more complete
person."
She said the AI-powered Plato would boost analytic skills
and problem solving, adapting in real time to a student's pace, prior
knowledge, and even emotional state.
The summit was notable: it brought together representatives
from more than 40 countries — including Olena Zelenska, Brigitte Macron and
Sara Netanyahu — alongside major tech companies such as Microsoft, Google and
OpenAI.
The Inversion: What Could Go Wrong?
This is where applying Munger's inversion is genuinely
important. The stated vision is personalized, patient, always-available
education. The dystopian inversion requires asking: what are the structural
incentives of the entities building this, and who controls what the robot
teaches?
Problem 1: Who writes the curriculum? A human teacher
is accountable to a school board, union, parents, and peers. An AI robot
educator is accountable to its manufacturer and whoever holds the software
license. President Trump appointed Meta CEO Mark Zuckerberg, Oracle's Larry
Ellison and Nvidia's Jensen Huang to a council that will weigh in on AI policy
— the very same day Melania's summit was promoting AI in children's education.
The overlap between the companies building the robots, the companies advising
on AI policy, and the government promoting adoption in homes creates a feedback
loop with no independent oversight.
Problem 2: The optimization target problem We just
spent the last section establishing that Meta and YouTube's algorithms —
optimized for engagement — caused depression, anxiety, and body dysmorphia in
children. An AI educator optimized for "personalized learning
outcomes" could just as easily be optimized for engagement,
time-on-device, or data collection. The incentive structures don't
automatically change because the product is called "educational."
Problem 3: The socialization deficit Teachers union
president Randi Weingarten strongly pushed back, arguing that AI is a tool
requiring human oversight and that education and decision-making should not be
delegated to the technology, asking: "What are we going to do to make sure
that AI is a tool? That the human beings are in charge, not the tool?"
Children learn social negotiation, conflict resolution, empathy, and
frustration tolerance from human relationships — including imperfect teachers.
A patient, infinitely available robot that never has a bad day removes the
friction that builds resilience.
Problem 4: The data harvesting dimension A robot in
every home, adaptive to a child's "emotional state," is also the most
intimate surveillance apparatus ever built. It knows when a child is
frustrated, curious, bored, or distressed. That dataset — even if never
"sold" — is enormously valuable for advertising, political targeting,
and behavioral prediction. The same companies that just lost $381 million in
lawsuits for exploiting children's psychology on 2D screens are now being
invited into children's bedrooms in three dimensions.
Problem 5: The worst-case Munger inversion — centralized
ideological control at scale This is the one that genuinely resembles a
sci-fi scenario. If a single AI system — or a small number of systems from a
handful of companies — becomes the primary educator for millions of children
across 45 nations, whoever controls the system's values, biases, and content
filters controls what the next generation believes about history, politics,
science, and identity. This isn't hypothetical — it's the same concern people
have about textbook publishers, but with personalization, emotional attunement,
and 24/7 access. The robot knows your child's vulnerabilities in a way no
textbook ever could.
The Paradox at the Heart of This Week
Here's the tension that makes this week historically
strange: On Tuesday and Wednesday, juries ruled that tech companies'
algorithmic systems caused measurable psychological harm to children and
companies knew it and hid it. On Wednesday afternoon, in the same news cycle,
the First Lady of the United States stood before 45 nations and proposed
putting those same companies' technology — now embodied in a humanoid robot — into
every child's home as a primary educator.
Brigitte Macron, who was present at the summit, touted
France's moves to restrict screen time and social media for children — a sharp
contrast with the host's vision. That tension wasn't resolved. It was performed
and then the robot walked out of the room.
The Most Likely 12-24 Month Trajectory
On the legal front: Expect a wave of state-level
settlements as Meta and Google calculate the cost of continued trials vs.
paying out. The school district trials are the real danger — they're better
funded, have institutional memory, and the discovery process will produce more
internal documents. A big-school-district verdict could trigger the kind of
master settlement negotiation that Big Tobacco faced. Realistically, total
liability exposure over 5 years is in the tens of billions, though appeals
will drag this out.
On the AI education front: The vision is real and
accelerating, but the regulatory framework is nearly nonexistent. The most
dangerous near-term scenario isn't a rogue robot — it's quiet, normalized
deployment of systems with no independent curriculum oversight, no emotional
safeguard standards, and no liability framework — precisely at the moment
when we've just proven in court that these companies' prior "safe for
children" claims were false.
The same lawyers smelling blood in the water on social media
addiction are watching the AI education space very closely. The first AI robot
that's shown to have caused measurable harm to a child — whether through
harmful content, emotional manipulation, or social isolation — will trigger the
next wave of litigation. The question is whether regulation gets there first,
or whether we need another decade of harm and another thousand lawsuits to
force accountability.
No comments:
Post a Comment
Thank you!