R Street Institute Fellow, Tech & Innovation. Opinions and puns are my own. @jgwithrow@techpolicy.social
Washington, DC
Joined April 2011
- Tweets7.2K
- Following607
- Followers1.2K
- Likes4.8K
Josh Withrow retweeted
.@dr_inserra and I have a new paper out today. In it we look at youth online safety policy such as age verfication about the 6 clear potential harms/costs that should be weighed against any purported benefits. You can read the full paper here: cato.org/briefing-paper/onli…
Josh Withrow retweeted
Digging into the details of Georgia's social media age verification law in the aftermath of the Eleventh Circuit opinion last week.
I think I've decided part of the mad dash to pass these bills is to do so while legislators are primarily Gen X and older - most of them have been in government/politics/policy their entire careers - meaning they don't have tech background, many need their grandkids to help them with their iPhones.
Case in point: younger people will have to ask AI - "what is facsimile". Absolutely no one under 70 uses a fax machine anymore.
The screenshot shows the text for the parental consent requirement and how it is to be obtained. Calling a toll free number or doing a videoconference is completely unrealistic. Not to mention, you'll be talking to someone in India if the platform uses this option. They could also request the parent's government issued ID or financial or payment information - but don't worry, they're going to delete it.
The most likely option is the "commercially reasonable method" which will involve a third-party age verification/relationship verification provider. These groups have been lobbying hard for these laws. Big business for AV providers.
The verified parental consent in these laws is equally, perhaps more disturbing than the age verification itself because it requires more identity documents that prove not just who you are but your relationship to the minor child...moving us from digital ID to a government family registry.
Josh Withrow retweeted
The concept that a full social media ban for kids requiring age verification for everyone "doesn't burden substantially more speech than necessary" is absurd. In what world is having to show ID just to go on social media not a substantial burden?
thehill.com/policy/technolog…
Josh Withrow retweeted
This is what encouraging learned helplessness in parents look like. Parents already control their children's access to a variety of dangerous scenarios. Loaded guns, dangerous drugs, driving vehicles w/out a license. This is not complicated. The govt makes a lousy parent.
Evidence on the harmful cognitive effects of TikTok, Reels, and other short form video keeps growing. It's not just the content, it's the attention fragmentation and the declining ability to plan for the future.
From @juliejargon
wsj.com/tech/personal-tech/t…
Josh Withrow retweeted
Haidt is right to raise concerns about how some kids use social media. But this new study is not the slam dunk for government intervention that some will make it out to be.
It followed adolescents in China, was observational rather than experimental, and found relationships that ran in both directions. Kids with weaker academic resilience or future orientation may also be more prone to compulsive short-form video use. That matters because correlation is not causation, and results from Chinese adolescents should not automatically be treated as proof of what is happening to American kids.
Even if some children are genuinely harmed by TikTok, Reels, or Shorts, the first question should be: where are the parents?
Parents can say no. Take away the phone. Delete the app. Set limits. Keep devices out of bedrooms. Different kids need different rules.
The answer is more parental responsibility and less political control.
Empower parents, not bureaucrats.
Evidence on the harmful cognitive effects of TikTok, Reels, and other short form video keeps growing. It's not just the content, it's the attention fragmentation and the declining ability to plan for the future.
From @juliejargon
wsj.com/tech/personal-tech/t…
Josh Withrow retweeted
On the AI policy point, Tyler gets things very wrong. He says we should be willing to cut the EA community some slack because “it would be a good thing if we paid more attention to AI issues, and to AI safety in particular.” ... “Score another point for effective altruism, at least if you make marginal rather than fully extreme adjustments.”
Excuse me, what? That’s really quite a jump. Plenty of people (including me) could agree it is worth paying more attention to AI safety, but we disagree strongly with the approaches many in the EA movement are proposing to address it. The problem is that a great many people and organizations in the EA movement who work on AI policy absolutely ARE talking about “fully extreme adjustments,” as Tyler calls them. They are pushing some of the most extreme and destructive interventions in the history of our republic all in the name of “alignment,” however they choose to define and impose it. And they are spreading mass panic to achieve that result.
Tyler is at least willing to admit that, “Many connected to effective altruism may exaggerate its effects, or perceive a high risk of doom without sufficient evidence.” But then he just gives them a complete pass on those antics. THEY ARE MAKING UP NUMBERS! All this “p(doom)” stuff is all just moralizing masquerading as science or “feelings dressed up as numbers,” as computer scientists Arvind Narayanan and Sayash Kapoor have noted.
I just published a long essay last night addressing both of these specific points and explaining why The Economist was right to call into question a movement that seeks to convert utilitarianism into a quasi-religion and treat morality like a math problem to be solved through elitist, and sometimes “fully extreme adjustments.”
Even if one thinks EAs may be “directionally correct,” how they propose to achieve their goals matters deeply. The ends do not justify the means.
adamthierer.substack.com/p/o…
Josh Withrow retweeted
"Free minds and parental guidance are the best means to protect children online." 💯
Thank you, .@RandPaul
Senator .@RandPaul blocked a shortcut to pass the Kids Online Safety Act (KOSA) without a debate or a recorded vote.
KOSA is sold as "child safety."
The fine print is a censorship bill: platforms decide what kids may see, and the only way to sort kids from adults is to check everyone.
Parents are the firewall.
A government mandated duty of care is not.
Thank you, Senator Paul!
Josh Withrow retweeted
Google’s latest reply brief in the search antitrust appeal presses a consequential question: When does winning competition for distribution become unlawful exclusion, and how should courts distinguish the gains from superior products from those attributable to unlawful conduct?
Josh Withrow retweeted
In @techpolicypress, I explain how age verification laws are opening the door to another AI cybersecurity issue: mandates are leading consumers to hand over personal information at the same time the barriers to cyberattacks are falling.
States should respond by vetting their existing laws for impediments to cybersecurity and with autonomous, AI-powered cyber breach detection and defense.
But the best method is to not pass laws the encourage the data to exchange hands in the first place.
Josh Withrow retweeted
The site blocking debate is not going great. Extremely frustrating to see lawmakers and lobbyists take a free and open internet for granted
"Applying blocking orders to DNS resolvers 'means that one court can cut off access to a website globally, based on a single individual’s filing and an expedited procedure.'"
Josh Withrow retweeted
These senators pay lip service to the Constitution but repeatedly work to undermine our American republic.
Protect free speech.
Oppose KOSA.
Josh Withrow retweeted
KOSA could come up in the Senate today with @MarshaBlackburn, and that makes the underlying question urgent: should Washington regulate kids’ online lives more, or should we empower parents?
The evidence on social media and youth mental health is much more mixed than the political rhetoric suggests. Much of the research shows correlation, not clear causation. Some kids clearly struggle with compulsive use, while others benefit from connection, information, and community. But the push for government control misses the trade-offs.
KOSA would still expand federal authority over platform design through a “duty of care.” The current Senate text does not literally require universal ID checks, but the broader age-assurance push raises real privacy and digital-ID concerns.
Parents can limit screen time, remove apps, monitor and stop use, and teach responsibility. Government should target predators, traffickers, fraudsters, and actual criminal conduct; otherwise, politicians and bureaucrats should get out of the way of parenting.
Stop KOSA. Less fear. More evidence. More parental responsibility. Less government getting between parents and kids. Empower parents, not bureaucrats.
Josh Withrow retweeted
what a sloppy piece of fiction from this Vanity Fair "journalist." I think she might just be a little miffed that I called the protestors she brought along "freaks." 😭
But I appreciate the pic they ran of me rallying the crowd. I'll happily lead this pro-progress revolution!
I FIGHT FOR FREEDOM!
Josh Withrow retweeted
Data After Dark is in Vanity Fair tonight
I sat for an hour-long, substantive interview as a co-host on our optimistic message. None was quoted. My read: they hoped we’d fumble and couldn’t spin it
Instead, they made it about race, sex, and a critical activist having a bad time
Josh Withrow retweeted
The @LawEconCenter submitted comments to the FTC on personalized pricing.
It was really a weird policy statement.
I don't know how else to say it (besides spending 40 pages going through it in detail). A few examples:
The FTC says the effects of personalized pricing on consumers are "unclear." It also says failing to disclose it is "likely" illegal.
The Statement relies on six economics papers to support its claim that more sophisticated personalization is less likely to benefit consumers.
None tests disclosure. None supports the claim as stated. Just a bunch of red herrings.
In all three empirical papers it cites, most consumers pay less or get lower offers under personalization. In two, the market grows.
That never comes up. Seems relevant to the topic and to the citations.
Dube and Misra is cited in every economic footnote. Great paper. I cite it all the time. The authors say outright it's not about competitive markets.
This isn't some detail. Competition is crucial for understanding price discrimination. That's been well known since Thisse and Vives (1988) or Holmes (1989). I have a paper on this. briancalbrecht.com/albrecht_…
Yet, it's being used to set disclosure policy for retail shelf prices.
I know this is a hot area. There's lots of pressure to do something. The states are doing dumb stuff.
But can we stop and think, study a bit? If there is a comparative advantage of the FTC it is deep market research.
laweconcenter.org/resources/…
The FTC proposed a new enforcement policy that would presume many uses of personalized algorithmic pricing to be unfair or deceptive under Section 5 of the FTC Act. I submitted a comment on how I think this proposal misses the mark. 1/
I agree with the argument made by George Ford at @lawandeconomics: “While personalized pricing may change who bears the cost of finding a better price, that does not make the offered price deceptive or unfair." phoenix-center.org/perspecti…
I also concur with @JuanMLondonoR and Chris Gardner at @CatoInstitute that the FTC is stretching its statutory authority under Section 5 with these guidelines. cato.org/public-comments/pub…
Josh Withrow retweeted
This junior engineer’s tweet has gone super viral claiming AI could kill us all. Few details of how! But I reviewed a book claiming exactly that by two of the most prominent “AI safety” advocates.
If that book is the best case for the Doomer position, I’m not worried. 1/
