The case for treating tech companies like railroads

Presented by

With help from Christine Mui

A leading critic of Big Tech has an idea for radically reinventing the relationship between Washington and Silicon Valley.

Writing in POLITICO Magazine today, Ganesh Sitaraman argues the dilemma of government’s role in platform moderation — touching on hot-button topics like political speech, child safety and AI’s risk to elections — can be settled by using a decidedly old-school set of policy tools.

“For hundreds of years, American law has been grappling with the problem of whether, when, and how to deplatform individuals and content… in public utilities and other infrastructural businesses,” Sitaraman writes. “Laws in these areas offer important insights into how to regulate social media and other tech platforms. The first step is to start thinking of technology a bit differently, less as a product or service and more as a utility.”

Getting classified as a public utility might seem horrifying to tech companies, which think of their services as premium, high-margin, extremely profitable brands — not commodities, like electricity or right-of-way. Additionally, any serious effort to move this direction in Washington would be met with overwhelming lobbying muscle. But his argument brings up a few key points for anyone, on either side of this debate, thinking seriously about the growing role of tech in society. We have faced challenges like this before; industries do transition from one social role to another; the law shouldn’t ignore that.

In a casebook published in 2022, Sitaraman and a group of co-authors argued that social media platforms can be classified as “networks, platforms, and utilities” (NPUs) that are subject to different regulations than traditional businesses providing goods and services. He writes that like the railroad, telecom and airline industries, today’s tech platforms provide a service so essential to “modern commerce, communications, and civic connections” that they demand specific regulation.

Sitaraman points by way of example to an 1881 case in which a carriage company sued its local telephone company for refusing to install telephone service due to its ownership of a competing carriage company. A Kentucky court ruled the telephone company was “bound to serve the general public … on reasonable terms, with impartiality,” helping set the precedent for a group of industries deemed too important to make their own, wholly financially-driven decisions.

This status didn’t totally handcuff the companies. Crucially for our current situation, companies were still allowed to refuse service in some cases: “NPUs were also allowed to exclude some customers, so long as the exclusion was reasonable,” Sitaraman writes, citing an 1839 court decision that innkeepers were allowed to bar disruptive customers.

“What history shows is that deplatforming is an endemic issue for any network, platform, or utility — it is not a challenge unique to social media or even tech platforms,” Sitaraman writes. “The reason is simple: For any network, platform, or utility enterprise, it is unworkable to serve literally everyone without exception because there will always be some bad actors.”

Deciding exactly who those bad actors are is where the discernment of courts and regulators enters the picture, in Sitaraman and his colleagues’ formulation. He posits a fine balance between NPUs’ basic responsibilities and their discernment in giving people the boot, based on legal liability — if a company can be held liable both for barring people unfairly or allowing them to be harmed by bad actors, the fulcrum will ultimately settle in favor of a fair, safe environment for users.

And what about just wiping out Section 230? Sitaraman warns that one commonly proposed remedy — removing Section 230 protection for liability altogether — might not work given the massive scale of tech platforms and their reach.

“Unlike the small amount of deplatforming that took place with telephones or train cars, tech platforms have millions of users and millions more posts. The 19th-century approach also places courts and judges at the center of the action, which could result in a patchwork of potentially idiosyncratic rules across the country,” Sitaraman writes.

But the other most commonly proposed solution of federal legislation is problematic as well, he adds, given the polarized political environment. Ultimately, Sitaraman’s argument for classifying Big Tech platforms as NPUs isn’t a solution in whole, but the beginning of one: Radically rethinking how government relates to the platforms that increasingly govern our public sphere, before settling on what the desired outcomes might be.

tech aftershocks in taiwan

The strongest earthquake to hit Taiwan in a quarter-century is a reality check about the risks of relying on one small geographical location to manufacture so many of the world’s microchips.

Taiwan Semiconductor Manufacturing Company, the world’s largest chipmaker that produces some 90 percent of the tech industry’s most advanced semiconductors, evacuated its factories Wednesday after the deadly 7.4-magnitude earthquake.

Jason Hsu, a former Taiwanese legislator, told DFD the disaster once again, shows the vulnerabilities of such a concentrated supply chain, especially if aftershocks spoil chips or continue to disrupt production.

“It’s a wake-up call,” said Hsu, now a senior fellow at Harvard Kennedy School. “You could sit around and wait for another earthquake, but then still have the same complaints.” Instead, he’s calling on the Biden administration to hasten its support for TSMC to open new fabs elsewhere, including Arizona, where it’s in the running for federal money from the CHIPS and Science Act: “Demonstrate a resolve to the world that you’re serious about this CHIPS business.”

TSMC, whose major clients include Nvidia and Apple, is already under political pressure to geographically diversify, given its quake-prone home and security concerns tied to neighboring China, which has threatened to absorb the island. It opened a fab in Japan last month, a second is in the works there, and two more are planned in Phoenix. The company says it’s still evaluating the earthquakes’ impact, but already expects to resume production overnight. Initial inspections found damage to some tools — none to construction sites or its critical high-end lithography machines, used to print cutting-edge chips.

Disruptions from a natural disaster are not unheard of. A 2011 flood in Thailand caused a global shortage of hard drives that lasted for months, and TSMC created new protective protocols after an earthquake hit Taiwan in 1999. — Christine Mui

more ai wariness

New polling from the Artificial Intelligence Policy Institute shows voters are wary of AI’s worst impulses.

In the poll of more than 1,100 respondents, shared exclusively with Digital Future Daily ahead of its publication, people largely say they’re willing to forego some of the benefits of AI as long as it avoids major risks. In a hypothetical scenario where AI is “able to generate beautiful works of art but has the capability to produce child pornography,” 63 percent say it should never be developed. For the less horrifying, but still bad scenario where an AI capable of producing “elegant writing” can also send spam emails, 51 percent say it should not be developed.

Additionally, the pollsters ran one experiment that seems to measure how the proliferation of AI has already caused people to question the content they see featuring public officials. “If audio comes out with Joe Biden’s voice supposedly accepting a bribe,” they write, roughly two-thirds of respondents said they would believe it was AI-generated, with one-third believing it was real. For the same scenario with Donald Trump, 58 percent of voters would believe it was AI-generated and 42 percent that it’s real.

Of course, neither figure can compare to the real leader of the free world when it comes to trust: “If a video comes out showing Taylor Swift calling her fans ‘gullible idiots,’” the pollsters write, just over three-quarters of respondents would believe it was fake and one-quarter that it was real.

tenderness toward the block

Wall Street is warming up to a tokenized future.

POLITICO’s Morning Money reported today on the growing ardor among investors for blockchain technology, which proponents say can make the transfer of real-world assets safer, more transparent and more secure. BlackRock’s first tokenized fund, BUIDL, is hosted on the Ethereum network and has attracted more than $280 million from large investors since it launched in late March.

“The whole idea is that blockchain as a public distributed ledger is a better way of representing ownership of things,” Securitize CEO Carlos Domingo told MM. His firm is partnering with BlackRock to run BUIDL, and he pitched tokenized financial products as an alternative for people who struggle to access traditional banking services.

“We’re solving for that by allowing [decentralized autonomous organizations], foundations and all these structures that exist in the crypto space to be able to access their account with us,” he said. “Some of these companies want to be transparent with what they do with their funds. They want to have transparency toward where their assets are. This is on-chain. You can go to Etherscan and the public blockchain and see holdings there.”

Tweet of the Day

The Future in 5 links

Stay in touch with the whole team: Derek Robertson ([email protected]); Mohar Chatterjee ([email protected]); Steve Heuser ([email protected]); Nate Robson ([email protected]); Daniella Cheslow ([email protected]); and Christine Mui ([email protected]).

If you’ve had this newsletter forwarded to you, you can sign up and read our mission statement at the links provided.