A fight is underway for the economic future of the United States, and some of our own politicians are siding with global competitors over domestic interests.

Artificial intelligence—and the data center infrastructure that powers it—represents the defining economic opportunity of this generation. JPMorgan recently reported that in the first half of 2025, “AI-related capital expenditures” contributed 1.1% to GDP growth, more than consumer spending contributed to overall expansion.

The nation that builds the most robust AI infrastructure will anchor the industries, jobs and wealth creation of the next half-century. That nation should be the United States. The question is whether policymakers—and those who influence them—will allow that future to take shape.

A new report from the American Energy Institute should be required reading for members of Congress, state legislators and local officials. It documents more than $39 million in foreign funding—from billionaires and foundations in Switzerland, the United Kingdom and Denmark—flowing into a coordinated network of activist groups working to block U.S. data center construction.

These are not isolated local protests. They represent a manufactured opposition campaign funded from abroad, with the goal of shifting America’s AI future elsewhere.

Foreign interests, watching the United States on the cusp of further technological leadership, are spending tens of millions of dollars to ensure we stumble. Meanwhile, some elected officials—wittingly or not—are amplifying that effort by calling for moratoriums, imposing regulatory hurdles and lending credibility to what is presented as a grassroots movement.

Before calling for a pause on data center construction, policymakers should ask a simple question: Who benefits? Not American workers who would build and operate these facilities. Not communities that would attract investment and tax revenue. Not the broader economy that stands to gain from the next wave of industrial growth.

The beneficiaries are America’s competitors—countries eager to capture the economic activity and strategic advantages that hesitation leaves behind.

A deeper concern is that some organizations traditionally associated with free markets and limited government have aligned, intentionally or not, with this foreign-backed campaign. Framing opposition in terms of property rights, local control or energy concerns may sound principled, but it can serve the same end as foreign-funded efforts to stall development.

Advocates of free markets should be wary of any movement that, however inadvertently, advances the goals of foreign interests seeking to constrain U.S. economic growth.

The debate over data centers is not merely environmental; it is strategic. Either the United States builds the infrastructure needed to lead in artificial intelligence, or it cedes that ground. There is no neutral position. Delays do not stop development—they redirect it.

Every moratorium, lawsuit or zoning fight tied to foreign-funded activism risks pushing investment and capability elsewhere.

Policymakers owe the public transparency about who is shaping the opposition they choose to support. Americans for Public Trust has warned that foreign actors are exploiting disclosure loopholes to influence domestic debates—raising concerns about transparency, accountability and sovereignty.

The United States has earned its place at the forefront of AI. Maintaining that position will require the political will to resist well-funded foreign interference. America’s economic future should not be outsourced.

Today’s TechByte was written by Bartlett Cleland, Senior Research Fellow with the Institute for Policy Innovation

The tragedy unfolding this week at New York’s LaGuardia Airport hints at human error. We shouldn’t jump to conclusions based on early reports, but in this case we have audio of the air traffic controller saying, “I messed up.”

Nothing but complete sympathy here for everyone involved, including the air traffic controller. Air traffic control is recognized as one of the most difficult, stressful and pressure-filled jobs on offer. And by all accounts it’s been decades since our corps of air traffic controllers was fully staffed.

But we must acknowledge the reality of human error. In many tragedies, human error is the No. 1 cause:

• Traffic accidents: 90 to 95% are attributed to human error.
• Aviation accidents: 70 to 80% are attributed to human error.
• Maritime accidents: 75 to 96% are attributed to human error.
• Medical errors: Human error is a leading cause of death in medicine, including misdiagnosis, medication errors, and surgical mistakes.
• Nuclear power and major industrial accidents: Both the Chernobyl disaster and the Three Mile Island incident were attributed to human error.
• General industrial/workplace accidents: Most are attributed to human error.

Human error is also a major factor in cybersecurity breaches, IT system failures, construction accidents, structural failures, power grid failures, and even in other areas such as business failure, including risk management errors and so-called “fat finger” trading errors.

Human error is neither a character flaw nor a lack of intelligence or integrity. It’s just a known element of human nature. That’s why in tragic cases like a child left in a backseat by a distracted parent, the proper reaction is compassion, not judgment. We all make mistakes.

Though human error may be inevitable, we are creating tools to help us compensate.

Human error is a perfect place to implement tools like AI to protect us from our own mistakes and compensate for our frailties. Implementing AI could substantially contribute to our quality of life by reducing preventable tragedies.

It’s likely that implementation of AI could eliminate most accidents due to human error. Chips and software don’t get tired, don’t get distracted, and make decisions in millionths of a second that a human might take 5 or 10 seconds to make. Studies suggest that humans can keep track of 3 to 5 independent variables at the same time, while AI can manage orders of magnitude more.

Humans are attention-constrained, while AI is computation-constrained. And while we can’t do anything to improve the former, we’re reducing the computational constraints on AI every day.

After the air traffic controller said, “I messed up,” some kind soul immediately responded with “No, man, you did the best you could.” And that’s entirely possible. It’s likely advanced AI could have prevented the accident, both pilots could have returned to their families, and an air traffic controller could be able to sleep tonight.

Every major innovation seems to inevitably generate an infrastructure fight—generating public and political opposition right up until we can’t imagine life without it. Railroads divided farms and ranches. Telegraph lines “uglified” streetscapes. Highways carved through neighborhoods. Pipelines, cell towers, wind farms, transmission lines—each sparked a familiar chorus: too big, too ugly, too loud, too much. And yet, in the end, the public benefited tremendously because economic growth requires a growing infrastructure.

Today’s favorite target for such opprobrium is the data center.

Just a couple of years ago, policymakers competed to attract data centers. They were symbols of growth, high-value investment, better tax bases, and the digital backbone of modern commerce. Now many of those same policymakers speak as if data centers are some kind of social evil—because they consume power and water, and because the AI boom is making that demand visible. You can see the shift in the sudden wave of proposed moratoriums and “pause” bills rationalized by energy, water, and ratepayer concerns.

Let’s be clear about what a data center is: critical infrastructure. It is the physical home of the digital services we rely on for banking, ecommerce, logistics, emergency communications, healthcare records, education, and—yes—national security. If policymakers are serious about economic competitiveness, they should treat data centers the way they treat ports, rail spurs, and power plants: as necessary facilities that must be sited responsibly, not demonized reflexively.

Critics raise legitimate issues—grid capacity, water use, noise, backup generators, local land-use impacts. But the policy response should be responsible and data-driven, not reactionary.

Start with electricity. Data centers don’t “steal” power; they buy it—often at industrial rates with long-term contracts—and their presence as a predictable buyer can justify new generation and transmission investments that benefit everyone. A rapidly growing load is not a moral failing; it’s a planning challenge. The right answer is to build abundant energy—more natural gas, more nuclear, more renewables where they make sense, more transmission—so that families aren’t pitted against servers in a zero-sum political drama. Lawmakers across the country are already debating how to protect ratepayers while accommodating growth, which tells you the debate is about policy design, not existential threat.

On water: in most cases, data centers use closed-loop systems for cooling, so it’s like filling a swimming pool one time. Over time, data centers consume less water than a car wash. Policymakers can require transparency, recycling, and closed-loop systems rather than discouraging new investment. Indeed, multiple states are moving toward water-use reporting requirements—exactly the kind of targeted, pro-information approach that beats panic.

The broader point is this: America’s economy won’t continue to innovate and remain prosperous by making infrastructure buildout difficult. We stay prosperous by insisting on sensible rules, clear property rights, and predictable permitting—and by letting private capital build the backbone of the next economy, instead of letting political panic veto it.

To ensure freedom of movement and freedom of commerce across state lines, the Constitution grants the federal government the power to regulate commerce among the states. This power, granted under the Commerce Clause, essentially established the United States as a free trade zone, where states are prohibited from enacting laws, taxes or regulations that inhibit or create unnecessary friction on economic activity between the states.

Under the Articles of Confederation, the original governing document of our republic, the 13 states tended to enact harmful populist policies designed to protect their own farmers and businesses from competition from other states. State tariffs and other protectionist measures hindered interstate commerce and contributed to a nationwide economic downturn that, in turn, led to the convening of a Constitutional Convention to remedy the problem.

It remains critical today for the federal government to preempt the states from their natural tendency to impose popular rules and regulations that affect businesses and consumers in other states or create friction for interstate commerce. The principle remains the same, though the Internet and e-commerce present new challenges and temptations the states are finding hard to resist.

Anything that happens over data networks is inherently interstate. The old analog phone network could distinguish between local and long-distance calls, but packets fly around the world at the speed of light, even if you are emailing someone in the room next door. Technology has made more of commerce interstate, such as voice communications, data exchange, e-commerce, and data hosting. Allowing the states to regulate inherently interstate commerce will simply slow commerce, the adoption of technology, and our global competitiveness.

As federalists, we need to recognize that there are legitimate roles for local, state, and federal governance, and while the states have allowed the federal government far too much regulatory control of functions that should be left to the states, regulating interstate commerce isn’t one of them.

That’s why at the Institute for Policy Innovation (IPI) we have consistently advocated federal, not state regulation of internet taxation, social media speech regulation and age verification, online privacy, VoIP regulation, and now artificial intelligence (AI). It makes no sense for there to be 51 different regulatory regimes for a business that travels across state borders at the speed of light.

It's just as much a violation of our constitutional, federalist republican system for a state to regulate AI as it is for the federal government to regulate state education. The Commerce Clause limits federal authority to interstate commerce and empowers the federal government to prevent states from burdening interstate commerce. Those of us who believe in limited government should insist on both.

Today is Global Encryption Day, and it’s an important moment to reflect on a foundational element of our digital lives that too many take for granted: encryption.

Encryption is not just a technical layer of protection—it is a shield for individual liberty, economic growth, and national security in the digital age. At the Institute for Policy Innovation, we’ve long defended the right to strong, uncompromised encryption, and today that defense is more important than ever.

We in the United States are fortunate. Unlike in many other countries—where governments either prohibit strong encryption outright or require surveillance back doors—we enjoy robust legal protections that allow individuals and businesses to use strong encryption to secure their data and communications. That’s not the case in authoritarian regimes, where encryption is viewed as a threat to state control. The ability to encrypt our devices, messages, and online transactions is a freedom Americans should not take lightly—and must actively defend.

Unfortunately, there are still those in our own government who argue that encryption should be weakened to aid law enforcement or national security. These arguments typically invoke terrifying scenarios: terrorists plotting in secret, criminals hiding evidence behind locked phones. The implication is that if the government had access to encrypted data, we would all be safer.

But this is a dangerous and false trade-off.

First, from a purely technical standpoint, there is no such thing as a “safe” back door. A vulnerability that allows government access will inevitably be discovered and exploited by bad actors—cybercriminals, hostile foreign governments, and unreliable bureaucrats. Weakening encryption makes everyone less safe.

But even more importantly, it reflects a fundamental misunderstanding of our constitutional system. The purpose of the American government is not to maximize security—it is to protect individual liberty. Our Founders did not envision a government that promises total security in exchange for diminished liberty. Freedom always comes before security. That is the American idea.

Strong encryption empowers individuals. It protects our private communications, our financial information, our intellectual property, and our constitutional rights. It also underpins the trust that fuels our digital economy. Without it, e-commerce would stall, innovation would suffer, and privacy would be an illusion.

So on this Global Encryption Day, let’s recognize encryption for what it truly is: a necessity for a free and secure society. We must resist any effort to undermine it—no matter how well-intentioned—and stand firm in the belief that freedom is worth protecting, even when it’s inconvenient for government.

Ever watch a drama set back in the time of kings, where the court jester or someone in the king’s retinue said something that displeased the king, and they knew immediately that their days were numbered? How everyone lived in fear of displeasing the king?

One problem with kings is that, because their power is nearly unlimited, everyone is vulnerable to the king’s caprice, malice and resentments. If you want to be safe, you must take up all the king’s grievances and desires. The king is everyone’s customer.

Thus, the king’s court ends up comprised of those who constantly amplify his whims to gain his fickle favor.

That’s monarchy, not self-government. Our American political system was designed to prevent capricious rule by One. But something like that is happening right now, and the latest example is the Jimmy Kimmel saga. Let’s walk through it, in chronological order:

1. Donald Trump doesn’t like Jimmy Kimmel and threatened to get him pulled off the air.
2. Nexstar, which owns 32 ABC TV stations, announced just a month ago the purchase of TEGNA, another owner of stations. That transaction requires approval by the Federal Communications Commission (FCC). The Chairman of the FCC is Trump appointee Brendan Carr.
3. In its press release Nexstar committed to preserving “diversity of opinion.”
4. On Monday September 15, Jimmy Kimmel said something stupid about the murder of Charlie Kirk on his ABC show “Jimmy Kimmel Live!”
5. The following Wednesday FCC Chairman Carr said it is "really sort of past time that a lot of these licensed broadcasters themselves push back on Comcast and Disney, and say, 'Listen, we are going to preempt, we are not going to run, Kimmel anymore until you straighten this out, because we licensed broadcasters are running the possibility of fines or license revocations from the FCC if we continue to run content that ends up being a pattern of news distortion.'" Carr went on to say, "When you see stuff like this—I mean, we can do this the easy way or the hard way.”
6. Within hours, Nexstar, which requires Carr’s approval for their TEGNA purchase, followed Carr’s “suggestion” and preempted Kimmel. Sinclair soon followed.
7. Shortly thereafter, Disney/ABC announced the suspension of “Jimmy Kimmel Live!”

Now, where’s the problem? There’s nothing wrong with Jimmy Kimmel expressing his political opinions on his late-night TV show, “Jimmy Kimmel Live!” He has been given that platform by ABC.

There’s nothing wrong with Nexstar and Sinclair deciding to stop carrying “Jimmy Kimmel Live!” They own the stations.

There’s nothing wrong with Disney/ABC deciding to suspend “Jimmy Kimmel Live!” It’s their program.

But there is something very wrong with Chairman Carr using government power to pressure companies into appeasing the Trump administration.

Why? Because it’s a violation not of the First Amendment, and not of FCC authority, but of the principle of limited government.

The Founders created a system of limited government to protect Americans from the capricious rulership of flawed human rulers. Conservatives have always believed in limited government for this and other reasons.

For conservatives, limited government is both a principle and a norm, and norms matter. If we abandon that norm, both parties will rush to use government power to punish enemies and reward friends, and we’ll be right back to rule by capricious, flawed human beings.

It’s always a challenge in public policy when there is an apparent conflict between two reasonable positions. Eventually, these conflicts get worked out through the legislative or administrative process, but that can take a while at government speed, which could mean harm in the meantime. That’s why, for instance, we have injunctions—to prevent harm to a party while a dispute is being adjudicated.

A current dispute is the regulation of “artificial intelligence” (AI), better referred to as large language models (LLMs). These models include names you recognize, but also many more purpose-built models designed to perform specific business and industrial tasks much faster and more accurately than humans. 

The potential of LLMs should not be a surprise, as the value of a new tool is not a surprise. Already, many are using LLMs as valued assistants in their work and daily lives.

Of course, there are concerns about LLMs, about AI taking over the world, killing or harming humans, destroying jobs, embodying racism or antisemitism. But the concerns we most hear about are straight out of science fiction. (Yes, we’re aware of the recent Grok incident.)

An immediate concern is something much less sensational: Copyright. Those behind the LLMs have been quite cavalier about their use of materials protected by copyright to train their models. Already a court has found that use of copyright material for training LLMs is fair use, but unauthorized access to the material was piracy and subject to statutory damages. Other cases will of course follow.

Deciding how LLMs fit within our existing laws, or making new laws, is a process and will happen. But one thing that could bollix things up in the meantime is a patchwork of differing and conflicting state LLM regulations.

These technologies are inherently interstate in nature and thus should be regulated at the federal level as per the Commerce Clause. That’s why the penultimate version of the “One Big Beautiful Bill” contained a federal preemption of state AI regulation. But it was stripped out at the last minute at the demand of Sen. Marsha Blackburn, a fierce advocate for copyright. 

Making life difficult for AI companies with state regulation is harassment, not sound policymaking. It creates friction for AI innovation, which is what opponents of AI and “Big Tech” want.

Copyright is a federal matter, and AI is interstate commerce, so both should and will be determined at the federal level. Between Congress, the courts, the Federal Trade Commission and the Copyright Office, rules will be written, cases will be adjudicated, injunctions will be requested, and damages will be assessed.

We still need a federal moratorium on state AI regulation until Congress and the courts establish clear rules that protect property rights while otherwise not restricting AI innovation.

The House Energy & Commerce Committee was scheduled to “markup” the Kids Online Safety Act (KOSA) this morning but cancelled the markup at the last minute. (Proposed bills go to a committee markup to give members opportunities to change or amend a bill and to see if there is committee support to move the bill along toward a possible full vote.)
 
In the case of KOSA, there are too many problems to advance the bill, even though many proponents have the best of intentions. So it’s a good thing that the markup has run into problems.
 
Let’s focus on the biggest problem: Mandating age verification.
 
The best online privacy practices include the following principles: Don’t collect personally identifiable data (PID), such as driver’s license numbers, physical addresses, Social Security numbers, etc. And don’t store information any longer than necessary to provide the service, because of the cybersecurity danger of having information stolen and sold to criminals.
 
It isn’t a perfect system, but the most harmful hacking incidents have been when such information WAS collected and persistently stored.
 
Here’s the thing: KOSA requires that online platforms do ALL of those things. Companies MUST collect personally identifiable data, and they must store it.
 
And they won’t just be collecting information on minors. How does a platform know that a 45-year-old isn’t fourteen without performing age verification on the adult? In fact, there is no such thing as age verification—there is only identity verification, of which age is a subset.
 
So either the platforms themselves will have to verify street addresses against property tax databases, or verify driver’s license numbers and birthdates.
 
And turning to a “trusted third party” solves nothing. AU10TIX is an Israeli company that functions as a trusted third-party verifier of identity for companies like TikTok, Uber, Upwork, Coinbase (gulp) and X (formerly Twitter). And guess what? AU10TIX got hacked, exposing users driver’s licenses and other personal information.
 
The way to protect privacy is to NOT collect personally identifiable information, not for the federal government to mandate its collection.
 
KOSA is also problematic on First Amendment grounds, for creating an undefined “duty of care” for companies, for creating enormous potential financial liability for platforms, and for granting significant new powers to the administrative state in the form of the Federal Trade Commission. There’s a lot wrong with KOSA.
 
Just because the name of a bill or the intentions behind it sound attractive doesn’t mean that the proposed government solution isn’t worse than the problem. Congress should take a pass on KOSA until and unless the many problems with it can be resolved.

This morning, the Subcommittee on Innovation, Data and Commerce of the House Energy and Commerce Committee (whew!) held a hearing to consider two bills aimed at amping up federal regulation of Internet speech and privacy. 

• Kids Online Safety Act, or “KOSA,” is draft legislation intended to protect minors from alleged online harms.
   
• American Privacy Rights Act, or “APRA,” is a discussion draft (something less than draft legislation) intended to go beyond protecting minors to create a nationwide privacy standard regarding the way companies handle consumer data and privacy.

There has been a lot of discussion about the need for privacy legislation for years, but we think it was wise for governments to move slowly, if at all. Technology changes rapidly, innovators iterate frequent changes to their products, industries work to develop best practices, and consumers over time learn how to navigate ever-changing technologies.
 
Moving slowly and only legislating when needs have become clear and when there are demonstrated harms reflects a conservative approach to governing.
 
Of course, in our federal system, the states also legislate and regulate, and states have begun passing laws and regulations designed to address these same problems. But the states that have acted have passed laws that are in conflict with each other, have different definitions, and create differing compliance burdens for companies. Additionally, we expect many of these laws to be invalided through future Supreme Court decisions, because state legislators have been more concerned about scoring PR points than protecting the First Amendment, which is pretty clear about government regulation of private speech.
 
Further, and obviously, because the Internet is inTERstate rather than inTRAstate commerce, federal rather than state legislation is appropriate. The Commerce Clause is pretty clear about that as well.
 
So the APRA effort in particular could be useful in pre-empting conflicting state regulations and creating uniform national standards and definitions.
 
The problem is, while both KOSA and APRA have noteworthy intentions and useful provisions, neither of these bills is ready for prime time, and both contain significant problems.
 
KOSA is particularly problematic in that it empowers federal regulators to restrict the First Amendment speech rights of Americans and creates huge new legal liabilities in the process. KOSA would also give significant new powers to the Federal Trade Commission—an agency that is already out-of-control and grasping for more power during the Biden administration.
 
It’s possible that either or both KOSA and APRA can be refined through committee work to play a positive role in Americans’ privacy and data security, but we aren’t there yet. And, as always, just because the name of a bill sounds attractive and there are problems to address doesn’t mean that the proposed government solution isn’t worse than the problem. The Devil is always in the details, and there are still lots of problematic gremlins lurking in both KOSA and APRA.

Well, that didn’t take long. After the Biden administration finally secured the necessary commissioners to form an operating 3-2 Democrat majority, the Federal Communications Commission (FCC) announced Wednesday that it will hold a vote to again reclassify broadband networks under Title II of the Communications Act of 1934.
 
A brief history. Activists have been calling for federal regulation of the internet since the 1990s, but the Clinton and Bush administrations wisely took a light-touch approach, understanding that regulation would impose unintended consequences on this nascent technology.
 
As the broadband revolution began in the 2000s, the FCC chose to classify broadband networks as Title I information services rather than as Title II common carrier services. The difference was significant, because under Title II the FCC would have almost unlimited regulatory power over broadband.
 
We’ve all seen how broadband networks have thrived under a light-touch Title I regulatory authority, with the growth of ecommerce, streaming entertainment services, social media, and the smartphone revolution.
 
But that innovation and economic growth has not deterred activists who simply believe in federal control for the sake of federal control. And in 2015 the Obama FCC, absent any compelling data, voted to dramatically change course and reclassify broadband networks as Title II common carriers.
 
The ominous threat of heavy regulation had the consequence of discouraging private investment in broadband networks, which the FCC found to have declined by around $1 billion between 2014 and 2015.
 
Over the 20+-year history, broadband network were only classified as Title II common carriers for a brief two years, from 2015 until 2017. The Trump administration’s FCC returned broadband services to their original Title I classification.

Since then, broadband speeds have increased, and real broadband prices adjusted for inflation have fallen 12 percent since 2017.  In fact, broadband prices have remained mostly flat during a time when inflation has been driving up the cost of almost every other consumer service.

In other words, there is evidence that Title II classification caused a harmful drop in infrastructure investment, but not a single piece of evidence of any benefit from Title II, much less any evidence that consumers have been harmed since that two-year experiment was wisely reversed.

But now, here we go again. For purely ideological reasons, and without data or examples of harm, the Biden FCC is reasserting its power to regulate broadband networks under Title II.

Not only is such a move unnecessary and in fact contrary to evidence, but these major policy pendulum swings every time the White House changes hands cause uncertainty for broadband companies and investors, which prevents the kind of long-term planning and financing necessary for major infrastructure investments.

So, net neutrality is back. But the Institute for Policy Innovation (IPI) will be there pushing back against harmful regulation of the internet, just as we have since the internet was being called the “information superhighway.”

On the surface, a “responsive” government seems like a good idea.
 
But the Founders recognized the tendency of elected politicians to chase the approval of voters could lead to bad policy and the trampling of rights, which is why they designed a system of checks and balances to slow down the responsiveness of government.
 
The way I have often responded is “we don’t want a responsive government; we want a limited government.”
 
A “rush to legislate” is almost always a bad idea. It’s better to let new innovations and technologies play out, and then only regulate or legislate if discrete problems have become obvious. Otherwise, we run the risk of letting precaution get in the way of innovation.
 
We see this rush to legislate in many areas of public policy, but today perhaps the best example is the concerns and fears about artificial intelligence (AI). At both the state and federal levels, legislators are rushing to “get out in front” of the issue and show voters they are being responsive.
 
Of course, protecting ordinary Americans and public figures against abuses facilitated by generative AI is a worthy goal. But many of the so-called “digital replica” bills currently moving in state legislatures go too far in creating unintentional legal liability for innocent bystanders. They could also violate First Amendment free speech protections.
 
Furthermore, in most cases existing law is adequate to cover the majority of concerns. States and the federal government already have laws against deceptive trade practices, fraud and deception, and laws governing name, image and likeness (NIL). Misappropriating someone’s likeness is already punishable under existing law.
 
But that hasn’t stopped states from moving legislation that could, in some cases, make it illegal for you to use color correction on a photo of a performer that you post on social media. Tennessee has already passed its ELVIS Act, which thankfully was significantly amended before passage, to address these concerns and to narrowly target performers’ rights. But many state efforts are far broader, as is the proposed federal No AI Fraud Act.
 
Imagine the minefield these digital replica bills would have created for the movie “Forest Gump,” for instance, where actor Tom Hanks was digitally edited into a host of historical footages. Without First Amendment fair-use protections, much of the film might have been subject to legal liability under these new digital replica bills. The same could be said about the current streaming hit “For All Mankind,” which is an alternate history of the Space Age in which, for instance, John Lennon is not murdered and lives into the Reagan administration and is portrayed as criticizing President Reagan.
 
Again, it’s eminently reasonable to protect performers as well as ordinary Americans against abusive uses of AI. But legislators should be cautious in their rush to legislate on AI to not rule out tools that can be used for innovation and creativity, especially when existing law covers most concerns.

Yesterday, the United States Supreme Court heard oral arguments over state attempts to regulate content moderation on social media platforms, NetChoice v. Paxton (Texas) and Moody v. NetChoice (Florida). The two cases differ in some details but not in intention and certainly not in constitutional harms.
 
Driven by outrage over “Big Tech’s” supposed discrimination against conservatives, the very red states of Florida and Texas jammed legislation through that allowed state legislators to demagogue against the tech menace and to preen for the voters. But for many of us, the constitutional defects of the legislation were painfully obvious.
 
And those defects were made clear in Monday’s oral arguments. 

• Social media sites are private actors, not government actors. Only the government can censor speech, which is a violation of the First Amendment. Conversely, as a private actor, social media sites are protected by the First Amendment against government control of the expression they choose to host. The First Amendment exists to protect private actors against the government, not to empower government over private actors.
  
  
• Websites have a right to curate the experience they want their customers to have. One justice brought up the example of “shouldn’t a Catholic discussion website have the right to exclude onery Protestants?” A state regulation demanding viewpoint neutrality not only would bar that example but would actually require websites to allow antisemitic posts if they allow pro-semitic posts.

In a bizarre development, the Texas solicitor general cited the support of Tim Wu and Zephyr Teachout—two of the most radical progressive actors in the tech space—for the Texas legislation that was driven by social conservatives.

Which reminds us that social conservatives aren’t the only ones who want to regulate online speech. Empowering government to regulate online speech won’t be limited to the preferences of a couple of red states.

In other words, what is at stake is not just a few big social media platforms, but the future of online speech.

Ultimately, these efforts will fail. Most likely, the Supreme Court will find them to be unconstitutional. And over time the panic over social media harming society will take its place in the long line of panics over things like violent video games, rap music, racy television shows, pop music and even Elvis Presley.

But there is another scenario. The Texas law not only demands that social media sites enforce viewpoint neutrality, but it also outrageously bars such companies from leaving the Texas market. So if social media platforms are forced to comply with these laws, these sites would soon be filled with all sorts of harmful, offensive content, which would drive advertisers and users away from those platforms and eventually cause them to fail.

Unless that was the legislators’ real goal in the first place.

Both state and federal legislators are actively pushing legislation designed to “protect our children online.” Of course, there is already a federal Children’s Online Privacy Act that regulates the online collection of personal information about children under 13 years of age to such a degree that social media sites largely disallow those under 13 in the first place.
 
So, who we’re really talking about are teens, and the legislators who want government to step in and limit their time or completely stop them from accessing social media by requiring social media sites to implement age verification.
 
But these proposals go well beyond a self-declaration of age like one might find on a liquor company’s website. Rather, the legislation would impose legal liability and large fines on companies that do not block teens from opening a social media account, or from being online at all, unless they can prove parental consent.
 
Often the sponsors assert that “we” should know if “kids” are online. That’s fair enough since those enabling access to being online should definitely know if their teens are online and what they are doing. But rarely are proponents of age verification forthright about what’s involved in requiring age verification and knowing who is online.
 
There are moral, ethical, constitutional and even patriotic arguments against these age verification schemes but the operational challenges alone are worth noting. Because you cannot verify age without verifying identity, the first task a company must face is determining that an account is being set up for a teen.
 
Data and proof will have to be collected to verify the identity of the teen who wants the account and also for the adult who is required to approve. Evidence will be necessary to document the relationship between the teen and the adult, parent, guardian, teacher, etc. Proof that the adult has the authority to approve the teen will also be needed. Then consider the complications of divorce and custody, or teens in foster care. The documentation required will be extensive and intimate.
 
What about adults who simply want to open their own account? They will have to prove their identify to prove they are an adult. No more anonymity. Think that’s a great policy goal? Perhaps for those who aren’t victims of stalking or an abusive marriage, or people trying to start over while reconnecting with supportive friends.
 
How will a company avoid liability if the data is correct but the submitter is gaming the system? Perhaps a biometric identifier will be needed, eye scan or fingerprints on the device so that the person claiming to be an adult can be verified as a specific adult.
 
Of course, all the collected data will have to be stored/retained at least until a statute of limitations runs out just to protect the company from legal liability. Do we really want all that information about adults and their teens stored in a giant database?
 
These proposals seem non-serious, or perhaps the legislators and organizations peddling such intrusive government across the country hope people just won’t think through the details. The shift toward heavy regulation and liability might lead some to speculate whether the real motive is to drive up costs for social media so the companies will be put out of business by weaponized government.
 
Protecting kids is a worthy goal for us all, but there are ways to do it that don’t require sacrificing our liberties and violating our privacy, which is a harm in itself, including for our teens.

You can be excused for losing track of the many new federal spending programs (and their accompanying acronyms) that were summoned into existence by the Biden administration and a compliant Congress in 2021’s $1.2 trillion Infrastructure Investment and Jobs Act. You can have a lot of Big Government fun with a trillion or so dollars.

Most conservatives had little enthusiasm for the infrastructure bill, and even less in the ensuing years. Inflation quickly rose to above 9 percent 10 months after the infrastructure bill became law. Besides, most conservatives never really bought into the argument that the U.S. suffered from “crumbling infrastructure” that required huge taxpayer subsidies to remedy.

Communications networks are obviously infrastructure, so it was no surprise that $65 billion of the $1.2 trillion created several new programs designed to increase availability and access to broadband networks.

It’s hard for fiscal conservatives to justify $65 billion in taxpayer dollars to subsidize more broadband, especially since somewhere around 94 percent of Americans already have access to broadband through private sector providers. And, for the most part, we don’t.

But access to a physical broadband connection is different than adoption—actually subscribing to a broadband service. So as broadband networks became nearly ubiquitous, the “digital divide” discussion morphed from access to adoption. And there are a lot of American households who could connect to broadband but do not for various reasons.

Depending on whose numbers you use, something around 40 percent of low-income households do not subscribe to broadband services or have a computer in the home. And while this may seem like a matter of individual choice, the pandemic made it clear that access to broadband is a social good—not only because of the need for access to remote education, but also to telemedicine and other critical services.

So with $14.2 billion of the infrastructure bill, Congress created the Affordable Connectivity Program (ACP) to help subsidize broadband adoption by low-income households. To date, around 23 million households have gained household broadband because of the ACP.

One virtue of the ACP program is that it didn’t require the creation of a new bureaucracy or the hiring of a large number of new federal workers. The benefit goes directly to needy households. It’s also stingier than the pandemic-era program that it was designed to replace. All-in-all, as federal programs go, targeted and efficient.

But ACP funding is projected to run out in April, and this week the FCC began phasing down the program. The likelihood is that most of the 23 million households assisted by the ACP will once again lose broadband service.

With $6-7 billion, Congress could extend the ACP for another year, and there are various proposals to do so, but so far, none of them have made much progress.

IPI will be on-guard against wasteful state and federal spending on duplicative new taxpayer subsidized broadband projects that compete against existing private sector providers. Much of the infrastructure spending was unnecessary, much will be poorly allocated, and some will actually be counterproductive.

But the ACP program, by comparison, is an efficient and targeted program that accomplishes its intended purpose.

If Congress were of a mind to scrap all non-essential spending to set our fiscal house in order, we’d be ecstatic about that. But short of such conviction, it seems unwise to allow the ACP program to expire through congressional neglect.

The Texas foster care system is infamous for failing its wards. The system is so flawed that in 2015 it was found in violation of the 14^th Amendment, which guarantees the right for people to be free from an unreasonable risk of harm while in government custody. The reason: Kids entering the Texas system came out in worse shape than when they entered. Many kids go in because of physical abuse at home, so how bad is the system?
 
In 2022, more than 3,100 foster kids went missing. And every year around 40 kids die while in the care of Texas. What more will it take to get the state to take seriously its responsibility to children in its care?
 
Some changes were made this legislative session, but the state Legislature decided to move in the direction of laws to favor parents in child abuse investigations, hoping that this will reduce the number of kids entering state care. But at least a new mandate requires that the state provide foster kids duffel bags and backpacks to transport their belongings, instead of the trash bags they had been providing.
 
Enhanced parental rights and backpacks aside, some will still end up in foster care, and they deserve to be safe and secure, not made worse off. Utilizing modern technology and the private sector could save lives and improve care.
 
Texas did partner with the private sector, a tacit admission that the state was failing on its own. The community-based care system now being implemented is a collaboration model between the Departments of Family Protective Services and local private sector non-profit organizations.
 
But when it comes to using the best technology to protect kids the state simply fails. The current system, built 30 years ago, contributes to welfare services losing track of kids, because caseworkers routinely have to track down physical documents in an intensely paper-based system. As the federal court said, this system “creates opportunities for important safety-related tasks to ‘fall through the cracks.’” But eight years later, same system.
 
Yet the state has no plans to upgrade to something effective even when less expensive, easy options are available. Gone are the days when an expensive custom software build was necessary. Software-as-a-service provides a license to use software on a common platform that can still be molded to unique needs. But even these improvements would require the state to devote adequate resources, which has been a chronic problem.
 
Given the introduction of another state entity into the mix with the non-profit organization, this sort of flexible platform would also help safeguard the family and child information that should be made available electronically. A new system would also empower social workers with the best individualized data available to help them stay ahead of potential problems, not least to help end the ongoing crisis of lost kids.
 
If only a fraction of the attention that state legislators have paid to demonizing “Big Tech” was focused on where the state can directly help at risk kids using technology, much good could already have been done. Then again, populist demagoguing is easy and cheap, and caring for kids is hard and costs money.

Many outside of Texas enjoy mocking Texans with “I know, I know, everything’s bigger in Texas,” but the reality is that Texas is really pretty darn big. The US Census reports that as of 2021 Texas has more than 30 million people, living in more than 12 million households spread out across more than 261,000 square miles. As of the census, 87% have broadband at home.
 
More recently, BroadbandNow.com indicates that 92.5% of the Texas population has access to wired or fixed wireless broadband. Reaching that many people in a large and diverse state is no small feat, but the need, or value, for all Americans to have access to broadband became clear during the COVID-19 pandemic. Understanding that value motivated the U.S. Congress and the Texas state legislature to allocate billions of dollars to help ensure that all Texans have access to broadband.
 
Texas will receive $3.3 billion of federal funds made available from the Infrastructure Investment and Jobs Act specifically to expand access to broadband. This year, the Texas legislature took a half measure and allocated an additional $1.5 billion to expand internet availability across the state through a new Broadband Infrastructure Fund but only if voters approve it at the ballot box in November. This is all in addition to billions more via other federal funds of which Texas will get a share.
 
Exactly where the Infrastructure Act and state funds will go is still unknown. The Texas Broadband Development Office will allocate the funds and presumably the money will go to communities where it is most needed to extend access. But this won’t happen automatically.
 
Strong guardrails on allocation and spending must be in place to guarantee that access is in fact expanded, as opposed to money being spent to “overbuild” in areas where Texans already have access. Unfortunately, making sure that the most effective technology is deployed is one guardrail that has already been compromised.
 
The Texas legislature passed legislation making fiber a priority for investment. But in some areas fiber is not the right answer and cable or fixed wireless may be more appropriate. Regardless, legislative mandate of a particular technology is almost never a good idea. In this case the result will be a higher cost of fiber deployment and therefore fewer people served and money wasted.
 
Another distraction is the desire by some to lavish cash on electric utilities so they can build a “middle-mile,” and pad their bottom line. The folly of this idea is fully described by IPI in a recent op-ed.
 
This huge tranche of taxpayer dollars needs to be spent wisely and leveraged with private sector spending so that more Texans actually gain access to broadband. If bureaucrats and politicians get distracted with fanciful ideas and cronyism the losers will be those Texans who have yet to join the digital economy.

The IPI policy nerd holiday calendar may look a bit different than yours but days like World Wi-Fi Day, which is June 20th, and World IP Day, April 26th of every year, should be part of your awareness, too. Why? These two days celebrate two of the fundamental building blocks of the modern innovation economy. 

World Wi-Fi Day is a global recognition organized by the Wireless Broadband Alliance, designed to encourage the coordination of “cities, government bodies, fixed and mobile operators, technology vendors, and internet companies to bridge the digital divide that continues to exist in the world.” 

In simple terms, Wi-Fi is a wireless networking technology that allows the devices we use every day, from laptops to phones to printers and much more, to exchange information. That is, Wi-Fi allows all those devices to operate creating a network via a wireless router. As the uses for this network grow so too does the wi-fi economy, and the United States thrives on Wi-Fi. 

The US ranks amongst countries with the widest Wi-Fi adoption. According to a study by the Wi-Fi Alliance, the US has 33.5 million paid and 18.6 million free public Wi-Fi access points. Most know Wi-Fi best as their in-home access resource for networking a myriad of devices, which is not a surprise given that 85% of broadband subscribers have Wi-Fi at home. Whether in the home or not, the economic dimensions and opportunities for growth seem almost endless. 

Already valued at $995 billion, the total economic value of Wi-Fi is projected to rise to more than $1.5 trillion by 2025. Driving that increase in value is the expansion of its uses. With robotics and greater automation, artificial intelligence and the increased networking of more devices, Wi-Fi is woven through everything we do for work, learning and play. As the volume of use increases so too will the need for more spectrum to avoid congestion. This is where Congress needs to act. 

Congress has allowed the FCC’s spectrum auction authority to lapse. Under previous and wiser policy the spectrum pipeline kept flowing while also providing the opportunity to balance licensed, shared licensed and unlicensed spectrum. As of now, no spectrum is flowing, and the future of U.S. innovation is therefore at risk. 

People love being connected wherever they go, and enjoy having their things connected at home. If glitches due to Wi-Fi congestion become commonplace they will rightly look to blame this Congressional inaction. 

The key is finding the right spectrum balance. Licensed spectrum provides direct revenue to the federal coffers, but the value of unlicensed spectrum to the overall economy is also substantial. Economically, auction proceeds are a revenue center, but unlicensed spectrum contributes to economic growth which also benefits the federal treasury. Spectrum policy should focus on promoting innovation and providing the right spectrum mix for the country. 

Allowing FCC auction authority to lapse is bad enough. But to allow the situation to persist is inexcusable. Spectrum is incredibly valuable and should be treated as the terrific natural resource it is – the cornerstone of a vibrant innovation economy and future.

A variety of legislative proposals across the country attempt to limit “children’s” or minors’ ability to be online, to see certain content or contribute content, or to set up user accounts. Most enforce this by targeting the technology with heavy fines if the technology provider fails to police those who use their platform, use an app or gain access to the website. For those who are simply interested in blaming and attacking tech companies this seems like a great idea, but the implications reach right to our basic liberties.
 
Generally, the legislative proposals would require that a child have permission from a parent or guardian before they could open an account or access a website. Some would actually forbid children (the ages vary) from having an account at all. Regardless of the intentions of the advocates, with just a bit of critical thinking the problems are easily exposed.
 
Unless the platforms are required to perform any uniform identity verification, a minor first has to be honest about being a minor. If a minor simply lies about their age, they avoid the entire age verification gauntlet. If the minor is honest about their age, an adult must provide the necessary permission for the child to use the platform. Which means that the technology provider will have to verify that the adult is an adult, which necessitates identity verification, with some sort of proof like providing a birth certificate, driver’s license number, Social Security number, or property tax record. This information would have to be verified by the platform by checking public databases and stored to protect the platform from any future charges of a wrongdoing.
 
People without kids may think these proposals have very little to do with them. They would be wrong. The platform or service provider, acting at the behest of government, is now mandated to collect particularly personal information, which most still consider an egregious violation of online privacy. 
 
And keep in mind this is about engaging in communications, free speech or being able to find and interact with a community of your choosing. This is not about, for example, buying cigarettes or alcohol where the familiar check of identification is performed even if the items are delivered. Rather, this is about fundamental protected rights.
 
A better approach is for parents to be aware of and use the tools already provided to them at the service provider, hardware, platform and software levels. Instead of grandstanding about how we need government to usurp the role of parents, government could play a role in educating parents, guardians, teachers and influencers about how to be safe online. Then adults could actually be in control instead of being controlled.

Imagine you’re in a bar, or a restaurant, or a retail business. Suddenly, someone starts shouting offensive, racist garbage loud enough to affect other patrons.
 
Should the management have the right to kick the offender out? Or give them a warning? Or ignore them? Maybe that person is a regular. Should the management be allowed to bar that patron for a few months, or even permanently? Of course it should, and businesses occasionally do.
 
Sports and concert venues routinely ban fans who cross the line and become violent or threatening to those around them. That’s because the owners of these private businesses have the right to ensure that their customers have an enjoyable experience while they are there.
 
Guess what? That’s content moderation. Just as the bar owner has the right to warn, expel and or ban a customer who violates the business’s right to provide a good customer experience, so do social media platforms have the right to warn, block, or even ban users who use the platform in a manner that, in the management’s opinion, degrade the experience of other users.
 
This stuff really isn’t that hard.
 
Let’s go back to our original scenario. Should the bar or restaurant be held legally liable for the offensive statements of the rowdy patron? Again, of course not. The business didn’t engage in offensive speech—one of its users did. It’s the offender who should bear any legitimate liability, not the unfortunate owner of the retail establishment.
 
What if a murder plot was hatched in the restaurant? Is the restaurant criminally liable? Of course not. The restaurant isn’t responsible—those who executed the plot are liable.
 
What the much-discussed Section 230 of the Communications Decency Act does is simply extend the same kind of reasonable protections that already exist in the analog world to websites. Facebook’s website, Twitter’s feed, etc., are really no different than retail establishments where people gather and socialize or do business. Sometimes someone has to be warned, or suspended, or banned for their conduct.
 
Section 230 gives websites the ability to moderate content without being legally liable for such moderation, and it protects websites from liability for users’ conduct on those websites. It’s not some kind of novel or atypical liability shield—it’s a logical extension of the kinds of protections that exist in the analog world to the digital world.
 
Yesterday the Supreme Court heard oral arguments in which a party is attempting to hold Google legally liable because radical Islamists managed to get their videos uploaded to YouTube. The family of a woman who was killed in an Isis terrorist attack is attempting to hold Google legally liable because of the presence of those videos, and claims that YouTube’s algorithm promoted those videos based on user interests.
 
This strikes directly at the purpose and function of Section 230. But it also strikes at the concept of responsibility. Is the killer responsible? Yes. Are the killer’s accomplices liable? Yes. Is YouTube? Of course not—not under any concept of responsibility as understood in the history of modern civilization.
 
You could lose a lot of money betting on Supreme Court outcomes based on oral argument, but the good news is that justices of all philosophical stripes expressed serious reservations about the plaintiff’s arguments. Let’s hope so, because making websites legally liable for the conduct of their users would be the end of user generated content and user participation on websites.