POLITICO’s weekly transatlantic tech newsletter for global technology elites and political influencers.

By MARK SCOTT

Send tips here | Subscribe for free | View in your browser

THIS IS DIGITAL BRIDGE, and I’m Mark Scott, POLITICO’s chief technology correspondent. Thankfully, I am not traveling this week (for once). This meme made this sci-fi-loving journalist chuckle.

Make it so, Number One (OK, I’ll stop):

— A flurry of digital rulemaking in the United States has placed national security at the forefront. That’s a mistake if you’re looking for good long-term policy.

— Everyone has a story about how artificial intelligence may destroy the world. Let’s meet a British activist betting on the technology’s brighter side.

— What do the European Union’s spate of investigations into social media giants tell us about how Brussels is implementing its Digital Services Act?

WHY NATSEC CONCERNS ARE A CRUTCH

LONGTIME READERS OF DIGITAL BRIDGE will know my overt skepticism toward the idea of Washington passing any form of digital legislation. But let’s give U.S. lawmakers and policymakers their due. Congress got its way on a law that would force ByteDance to sell TikTok (within a year). The White House gave an update on how implementing its executive order on artificial intelligence was going (pretty well, apparently). Politicians are steadily chipping away at another federal comprehensive privacy bill — in what represents the best chance of rulemaking in a generation. Government tech-focused subsidies, either for semiconductors or other critical technologies, are starting to flow. Put these all together (alongside legislative long shots like kids-focused online safety rules), and we’re starting to see real movement. Is it time I changed my tune?

The short answer is no. That’s not because U.S. politicians aren’t eagerly writing laws. It’s because the driving force behind Washington’s willingness to dip its toe into digital policymaking is fears — real or imagined — that China is about to eat its lunch. You can see that in how the de facto forced sale of TikTok, in part based on unsubstantiated concerns Beijing had access to Americans’ sensitive personal data, came quickly after separate proposed legislation to outlaw so-called data brokers, those who collect massive amounts of data to sell for commercial purposes, from selling such information to foreign adversaries (aka China). That bill was similarly included in the far-reaching foreign aid proposals that U.S. President Joe Biden signed on April 24.

Time after time, national security has fast-tracked digital rulemaking that once seemed stuck in partisan bickering. Even the American Privacy Rights Act, the most recent attempt to pass federal data protection rules, has a smell of “NatSec” about it. If the U.S. wants to flex its muscles on artificial intelligence — on the global stage to counter China, so the theory goes — and that can’t happen without an updated nationwide privacy regime at home. I’m still 60/40 against the APRA getting over the line before the end of 2024. But if you frame data protection via the prism of national security, then, suddenly, there are more backers for the proposals.

I could go on. The billions of dollars set aside in both the Chips Act and Inflation Reduction Act should be viewed as a China-style economic plan to promote high-tech production around semiconductors, electric vehicles and other gadgetry. Sure, that makes total sense for the world’s largest economy seeking to maintain its global title. But, again, the specter of China looms large. The fact Washington was willing to lean on allies like Japan and the Netherlands to impose export bans on next-generation semiconductor equipment from heading to China is telling. U.S. officials adamantly tell me no undue pressure was placed on Toyko or The Hague. Other Western officials, however, say an ultimatum was made: You’re with us, or you’re not.

That’s where I feel Washington is doing itself a disservice. Of course, national security is paramount, especially when dealing with an authoritarian rival willing to use every lever to gain an advantage. But where the U.S. and other Western democracies have a massive advantage is in the basic freedoms, principles and commitments to openness, competition and fundamental rights that, frankly, are anathema to Beijing, Moscow and other totalitarian regimes. In a world where the European Union, not the U.S., has become the go-to digital rulemaker, it’s often easy to forget that, once, Washington was the leader in espousing an open, fair and interoperable internet to the world. It wasn’t long ago, for instance, when the White House announced the so-called Declaration for the Future of the Internet to reaffirm those tenets.

Such principles become harder to manage when digital policymaking is framed almost exclusively via national security concerns. Those often make decisions black or white. You sell TikTok, or you allow (allegedly) Beijing to collect data on Americans. You stop selling chip equipment to China, or you’re in a diplomatic “time out.” It allows for quicker policy wins by forcing lawmakers into flag-waving exercises to show who’s more patriotic in their support of rushed legislation that undermines basic U.S. values. At the start of Russia’s invasion of Ukraine, for instance, American diplomats strongly argued against the EU’s sanctions that stopped Kremlin-backed media from broadcasting in the 27-country bloc. The reason: You need to allow free speech, even if it means your enemies promoting themselves to your citizens. By forcing the sale of TikTok in the name of national security, those claims of putting free speech above anything else start to look pretty hollow.

In some way, it’s understandable Washington would revert to type by prioritizing the emerging threat of China. Congress remains broken, despite the flurry of legislative activity. And national security, even in a divided Beltway, still pulls Democrats and Republicans together like nothing else. But I question where this now leaves U.S. leadership that once promoted an open, free-flowing and messy internet to the world. For me, those days look numbered — all in the name of beating China at its own game.

AN ALTERNATIVE TO AI DOOMMONGERING

SITTING IN A SUNNY LONDON PARK THIS WEEK— with the sound of kids at a nearby playground and commuters grabbing a drink at a local bar before heading home — Hannah O’Rourke did something that few can: She surprised me. The young progressive activist runs CampaignLab, a mostly volunteer organization whose goal, in part, is to use technology to promote new ways in how political parties and voters interact. For the 2024 election cycle, where the United Kingdom is expected to go to the polls in the fall, that means finding novel uses for artificial intelligence. I spend much of my time looking at the most harmful use cases. (Shameless plug: The next chapter of my three-part Bots and Ballots series drops on May 7.) So to find someone genuinely attracted to AI as a means to boost democratic interactions — and not for mere political gain — was refreshing.

“The great thing about AI is it takes a conversation or language and turns it into a query of a bigger dataset,” she told me, geekily, over a coffee in southeast London. “So in terms of actually allowing accessibility to datasets, information, making it more accessible to people, there are so many really good opportunities here.” That’s not something you hear many, especially on the left of politics, willing to say out loud. Yes, artificial intelligence may cause harm, especially around politics. But it can also be used to open it up in new ways. “That shouldn’t be lost at the expense of worrying about some of the very real threats,” O’Rourke added. “You’ve got to see it as a tool. And ultimately, people who want to do bad things will be using this tool,” she continued. “So the question is: How do we, as people who want to do good things, use this tool in a way that is in accordance with what we think is right?”

What does that mean in practice? CampaignLab, which started in 2018, runs monthly hackathons where mostly progressive coders, engineers and activists hash out solutions to either technical or policy questions. Some products arising from these meetups include an AI chatbot that allows activists heading out to campaign at people’s homes to practice their pitch. The organization built the door-knocking training bot to mimic different voters with different personalities, depending on a few prompts. Then activists can converse with the AI-powered bot to hone its skills. The discussion is then reinserted into the training model and the system spits out areas where the individual can improve. “There are some really interesting, creative kinds of AI solutions that place humans at the center and can help us be better at what we do.”

Ahead of the British election, O’Rourke and her team built another AI bot that pulls together information from the country’s Electoral Commission, the independent regulator overseeing the upcoming vote. Then, campaigners can ask specific questions, based on that data, to better understand what’s allowed under the law. Another project pulled together campaigning material from candidates to easily classify their positions on key issues. A fourth scraped publicly available information from political discussions on TikTok to see what people were really talking about.

“Don’t be afraid of the technology,” she told me. “It’s not going anywhere, so you need to know how to use it.” O’Rourke, who during the Covid-19 pandemic helped campaign against the use of complex algorithms by U.K. authorities to decide which high school students could go to college, separates AI uses into two categories: those that empower individuals to make better choices by opening up treasure troves of data, and those that take away people’s agency by assigning decision-making powers to such algorithms. “There are insights to be gathered from this data,” she said. “I think that’s very democratizing, actually, because it allows people more access to information.”

O’Rourke doesn’t hide her political affiliations. She is an active member of the country’s opposition Labour Party and all of her organization’s work is aimed at helping progressive causes. Still, CampaignLab’s projects are all open-source, so there’s no reason why conservative groups, if they so choose, couldn’t piggyback on its work. Once, as a trial, she and her colleagues trained an AI bot to mimic the beliefs and mannerisms of Nigel Farage, a right-wing politician most famous for his role in the U.K.’s Brexit referendum. “I got Nigel Farage to explain why low-traffic neighborhoods were a good idea,” she laughed. “And that was quite fun. He came up with some really good lines.”

For many following the AI policy debate, there is a hard line between those who see only the downside of the technology and those who favor a purely positive outlook. It’s rare to find someone openly willing to suggest that artificial intelligence, in itself, is not good or bad. It is, after all, a tool dependent on the person who uses it. “People ask me if I’m an AI doomer or AI optimist,” O’Rourke told me. “I’m definitely in the middle, probably.” That “probably” is doing a lot of heavy lifting in that sentence. But in a world where many want to focus on the (overblown) concerns around deepfakes or the potential existential threat that AI may pose to the world, the British campaigner offers an alternative — one where the technology is an opportunity to improve how politics actually gets done.

BY THE NUMBERS

SOCIAL MEDIA GIANTS IN THE SPOTLIGHT

AND SO THERE YOU HAVE IT. This week, the European Commission opened expansive investigations into how Meta (via Facebook and Instagram) was complying with the bloc’s new social media laws. The focus is primarily on how disinformation campaigns spread on these networks; how the company handled political content in people’s feeds; and (which is close to my heart) how Meta allowed outsiders access to its data in the name of transparency.

Those probes, which may not lead to fines or remedies, follow similar investigations into TikTok and X (formerly Twitter.) There’s also a lighter-touch probe into how generative AI may affect the upcoming European Parliament election in early June. Each inquiry is different. But the basic gist is this: Companies have pledged to uphold their new mandatory commitments — and executives have even praised Brussels for its new rulebook. Now, as enforcement kicks in, are these tech giants actually keeping to their word?

This inevitably is a “speed up, but go slow” process. There had been massive political pressure (looking at you, European Internal Market Commissioner Thierry Breton) to announce investigations — and now! But, for practical legal reasons, those probes need to follow due process, which means it’s going to be months before we have any answers. That may anger those hoping Europe’s new social media rules will bloody Big Tech’s nose. But if you’re going to stand up these cases in likely legal appeals, regulators can’t skip steps.

WONK OF THE WEEK

WE’RE IN UNCHARTED TERRITORY this week for WotW. The focus is on not one, but 22 high-profile individuals who were named to the U.S. Department of Homeland Security’s Artificial Intelligence Safety and Security Board.

Many of the names are well-known. OpenAI’s Sam Altman makes the cut. So do Sundar Pichai and Satya Nadella, the respective heads of Microsoft and Alphabet. A nerdy list of no-names boffins, this is not.

The new board is expected to help the American federal agency and others use “safe AI” in developing the country’s critical infrastructure. But given the corporate slant to the advisers, critics already claim it may lead to regulatory capture by Silicon Valley.

THEY SAID WHAT, NOW?

“It’s clear that companies are currently unable to make chatbots like ChatGPT comply with EU law, when processing data about individuals,” said Maartje de Graaf from noyb, a nonprofit that filed a privacy complaint against OpenAI under the bloc’s data protection rules. “If a system cannot produce accurate and transparent results, it cannot be used to generate data about individuals.”

**Our May forecast looks promising: POLITICO’s Energy & Climate UK Summit is coming up with exciting conversations all around the journey to net zero. While onsite registration has closed, there's still room for you online. Don't miss your chance and register today.**

WHAT I’M READING

— When should democratic governments use so-called deepfakes to promote their foreign policy objectives? Analysts at the Center for Strategic & International Studies weigh in.

— Europe’s new Artificial Intelligence Act overlaps significantly with the bloc’s existing data protection rules. Lawyers from DLA Piper break down how the two pieces of legislation interact.

— Google paused the rollout of new online advertising tools after British regulators raised both data protection and competition concerns related to the tech giant’s so-called privacy sandbox. More here.

— Trying to keep track of how Europe’s new digital competition rules are being enforced? Here’s a handy site from the European Commission that pulls all that information into one place.

— William Rinehart outlines the complexities of America’s policymaking response to TikTok in a report (disclaimer: from January) that’s definitely worth a read. More here from the Center for Growth and Opportunity.

— The Future Society pulled together eight recommendations for how best to govern so-called AI foundation models and generative AI. Take a look here.

SUBSCRIBE to the POLITICO newsletter family: Brussels Playbook | London Playbook | London Playbook PM | Playbook Paris | EU Election Playbook | Berlin Playbook | Global Playbook | POLITICO Confidential | Sunday Crunch | EU Influence | London Influence | Digital Bridge | China Watcher | Berlin Bulletin | D.C. Playbook | D.C. Influence | All our POLITICO Pro policy morning newsletters