The media in this post is not displayed to visitors. To view it, please log in.
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
BUCKLE UP. DIGITAL BRIDGE IS A SPICY ONE THIS WEEK. I’m Mark Scott, POLITICO’s chief technology correspondent, and as the world of politics appears to be going from bad to worse (no matter where you are living), here’s a reminder that what may appear as government overreach one day can quickly become standard practice the next.
Grab your beverage of choice and relax:
— Jose Fernandez, a senior U.S. State Department official, talks about transatlantic ties, artificial intelligence and China.
— Twitter took down more pro-Trump accounts than anti-Trump accounts. But does that equate to censorship?
— Brussels’s new economic security plan is a pitch to have a seat at the world’s top geopolitical table. It’s a little messy.
A VIEW FROM WASHINGTON
WHEN IT COMES TO FANCY TITLES, Jose Fernandez‘s is a doozy. Officially, he’s the U.S. State Department’s undersecretary for economic growth, energy and the environment — and is in charge of the agency’s myriad bureaus, including the one recently created for cyberspace and digital policy. The former mergers and acquisitions lawyer also is central to the revamped transatlantic relationship via the EU-U.S. Trade and Technology Council. So when I caught up with him last week, Fernandez wanted me to know that — like almost everyone I talk to within these transatlantic digital policymaking circles — ties between Brussels and Washington are stronger than ever.
“Two years ago, our relations with the (European Union) were mired in irritants,” he conceded. “We’ve been able to get beyond those irritants and we’ve expanded our goals to meet our ambition.” That sounds wonderful. But dig a little below the surface, and it’s easy to find a raft of tricky policy questions that still mire the United States’ relationship with one of its largest trading partners in difficulties. Case in point: the ongoing discussions over a so-called “critical raw materials” deal that would allow European automakers and their suppliers to tap into American subsidies baked via the U.S. Inflation Reduction Act.
Fernandez said those talks were ongoing (he wouldn’t go further than that). But, he added, the very fact senior U.S. and EU officials now sat down, in person, every six months was a sign of how important Joe Biden’s administration took the transatlantic relationship. “We have irritants,” the State Department official said. “You couldn’t have a $1.4 trillion economic relationship without them. But there’s a real commitment to discussing them even when we disagree and to try to work through them.”
OK, sure. But what about China? Despite broad consensus between the White House and the European Commission — the executive branches of each region/country’s political structure — the U.S. is more willing to push back against China compared with its European allies, according to multiple conversations with U.S. and EU officials. Not true, said Fernandez. “Both Secretary Blinken and his counterparts in the EU have said that we’re looking to de-risk from China, not de-couple from China, and on that, we are 100 percent aligned,” he told me. “I would not agree with that. I think you’re seeing very close alignment,” he added when pushed on the differences over China.
Turning to the next iteration of the Trade and Technology Council, which will be held somewhere in the U.S. at the end of 2023, what will likely be on the agenda? For Fernandez (who also declined to say where it would be held later this year), the focus will be on artificial intelligence; so-called export controls, or bans on certain goods from being shipped to countries like Russia; and additional U.S.-EU funding for developing economies as they roll out 5G telecommunications networks and upgrade their digital cybersecurity capabilities.
“There’s a desire to continue to work with third countries on 5G development,” he said. During the last transatlantic summit in May, both sides announced funding for Costa Rica and the Philippines, respectively, on such digital infrastructure projects. “AI, and in particular generative AI, presents unlimited opportunities. But at the same time, some of the consequences have to be managed.” Like almost all Western officials in recent months, those attached to the Trade and Technology Council are falling over themselves to be seen to be doing something (anything!) to corral the growth of services like OpenAI’s ChatGPT. That includes plans to create a voluntary code of practice, designed by Washington and Brussels, which companies can agree to follow to uphold basic principles like fairness, transparency and impartiality.
Yet what that code of practice — which is expected to be presented to other G7 leaders in the fall — will actually look like is still a work in progress. As of early June, it was merely a two-page policy memo, shared haphazardly between EU and U.S. policymakers. Fernandez, too, had little detail to give — and wouldn’t be drawn on the regulatory differences between Brussels and Washington on how to deal with artificial intelligence. “We are now in the process of starting, so I can’t really give you much info on that because there is not much,” he acknowledged when asked about what the voluntary code of conduct would actually entail.
ACCOUNTABILITY VS. CENSORSHIP, PART TWO
AS A LAPSED BUSINESS REPORTER, I LIKE NUMBERS. Figures — unlike, ahem, politicians, lobbyists and officials — don’t lie. So when I came across a research paper from Georgetown and Yale academics on how anti- and pro-Trump bots on Twitter performed (in terms of spreading partisan messages across the social network) in the build-up to the former U.S. president’s first impeachment trial in late 2019 and early 2020, my ears pricked up. It provides a snapshot of how such political social media users disseminate their messages online — and, more important, sheds light on accusations from some Republican politicians that Big Tech is unfairly censoring right-wing voices.
Before we get to the findings, let’s lay out the methodology. The researchers collected almost 68 million posts from 3.6 million Twitter users between December 2019 and March 2020. They broke those accounts down into potential bots, or social media users that posted at significantly higher rates than average people, and then did a content analysis, at scale, to see whether such automated accounts were more in favor or against Donald Trump. Finally, the academics tracked the types of content these partisan bots posted during the impeachment trial, and how those messages were spread more broadly across Twitter.
OK, the results — and, remember, you have to read this in the context of accusations that social media platforms have an unfair bias toward liberal voices. In terms of bots, though, it was an almost even split. Collectively, there were 10,145 anti-Trump bots (posting 9.2 million tweets over that time period) compared with 11,571 pro-Trump bots (with 9.8 million tweets). But even though these accounts represented roughly 1 percent of the overall Twitter users within the study, they posted 31 percent of the total content. Key takeaway: Twitter doesn’t have a significant bot problem, in terms of the number of accounts. But it does when you look at how those automated users dominate the online conversation.
“For the price you pay for them, it’s a good bang for your buck,” Tauhid Zaman, a Yale academic who co-authored the paper, in reference to the effectiveness of bots on Twitter. “Once they are up and running, the marginal cost of a bot is essentially free. So in information warfare, it’s a really efficient tool to use.” Yet the pro- and anti-Trump accounts weren’t uniform. The more liberal bots were more likely to post content from more mainstream media outlets, whereas conservative automated users more regularly circulated material from dubious sites, based on a non-partisan content analysis of the quality of links shared by these bots.
That had an interesting effect on Twitter’s response to these automated accounts at a time when the social media giant (in the pre-Elon Musk era and after he took over) has tried to clamp down (mostly, unsuccessfully) on bot activity. When Zaman went back, earlier this year, to see how many of the anti- and pro-Trump bots were still active on the Blue Bird, he found a marked difference. Since early 2020, Twitter had removed 40 percent of the conservative-leaning automated accounts versus just 12 percent of the liberal bots. For human partisan Twitter accounts, the same held true: 12 percent of Republican users had been removed by March 2023 compared with just 5 percent of Democratic accounts.
For some in the U.S. House of Representatives (looking at you, Representative Jim Jordan), that could be a smoking gun in the alleged censorship of conservative voters online. Right-wing Twitter accounts were four times more likely to be removed compared with left-wing Twitter accounts. But where this gets incredibly tricky is figuring out why Twitter removed the bots. Is it because the company hates conservatives? Almost certainly not (that’s how not companies work). Or is it because these accounts were statistically more likely to share either false or dubious information, which therefore put them at odds with Twitter’s terms of service that prohibit (or, at least, down-rank) the spread of such questionable material?
“How much you are a Republican correlates almost exactly to how much disinformation you share. It’s almost one-to-one mapping,” Zaman told me. “I can’t say Twitter is suspending you because you’re a Republican. They could be trying to enforce some general policy on disinformation. But the effect is that it looks like they are suspending Republicans more, in a biased way, versus Democrats.” Unpicking that relationship — between how much false content a partisan social media user shares, and how that puts an account in breach of platforms’ terms of service — is crucial to understanding why more right-wing users have been suspended versus their leftwing counterparts.
BY THE NUMBERS
EUROPE (KINDA) FLEXES ITS MUSCLES
THE EUROPEAN COMMISSION WANTS YOU TO KNOW it’s getting tough. In economic security proposals published this week, Brussels outlined how it would protect local companies from having their technology stolen by others; limit the ability of third-party countries (read: China) from using their economic muscle to strong-arm the 27-country bloc; and work collectively to stop authoritarian regimes from buying stakes (or the entirety) of European companies, particularly in high-growth areas like quantum computing and cybersecurity. “We also have to be clear-eyed about a world that has become more contested and geopolitical,” said Ursula von der Leyen, the Commission president.
You could already hear the cheers from Washington as Brussels began to follow its lead on these topics. But don’t count your chickens just yet. First, the word “China” — key to all of these geopolitical maneuvers — didn’t get a mention. That may sound childish to point out, but appearances matter. And the failure of the Commission to call a spade a spade on China’s economic global ambitions doesn’t bode well for the negotiations to come. (These are just proposals that must be hammered out with the EU’s other political institutions.) Second, let’s be clear: The EU is still divided on how robustly to push back against Beijing. Until those divisions are figured out, which is key as Brussels will have to work with national capitals to implement its economic security agenda, I wouldn’t count on the 27-country bloc to pull its weight on the global stage.
WONK OF THE WEEK
WE’RE STAYING WITH THE WORLD OF AI this week to focus on Anna Makanju, head of public policy for OpenAI. It’s not her first Big Tech job (if OpenAI now makes it into that elite club?). The New York-based executive previously worked as a senior policy manager at Meta, specializing in content regulation.
The former Biden aide also has a background that would make most of the United Nations look on with envy. Born in Russia to a Nigerian father and a Ukrainian mother, she moved to the U.S. in the early 1990s and studied linguistics at Western Washington University, from which she graduated at 16 years old. She later graduated from Stanford University’s law school.
“I’ve argued countless times that Putin excels at neither strategy nor tactics despite so much odd insistence to the contrary in the West,” she wrote on Twitter last year (her career includes a stint at the U.S. Department of Defense). “Unless the willingness to kill an unlimited number of people inside or outside your country to achieve a goal is strategic or tactical brilliance.”
THEY SAID WHAT, NOW?
“We need an all-hands-on-deck approach — because that’s what AI’s complexities and speed demands,” Chuck Schumer, the Democratic leader of the U.S. Senate, said when announcing his “SAFE Innovation Framework for Artificial Intelligence,” proposals to set guardrails, but not actual legislation, around this emerging technology. “Innovation must be our North Star. And it is SAFE innovation that we must seek.”
WHAT I’M READING
— Risks and harms from the online world will increase at an exponential rate, but the knowledge needed to combat those threats has been increasing within tech companies and across civil society, according to conclusions from a task force looking at trust and safety issues on the internet, overseen by the Atlantic Council.
— Google’s licenses for its “infotainment” services for automakers are not compatible with Germany’s revised antitrust rules that impose stricter obligations on dominant companies like the search giant, based on a decision from the country’s competition authority.
— A comprehensive review of popular messaging apps found a wide variety of security practices, as well as wholesale public misunderstanding about what was considered “encrypted,” argue Justin Hendrix, Cooper Quintin, Caroline Sinders, Leila Wylie Wagner, Tim Bernard, and Ami Mehta for Tech Policy Press. (Full report here.)
— Recent cyberattacks on government agencies and IT providers within Ukraine were directly tied to operatives, known as Cadet Blizzard, associated with Russia’s military intelligence, claims Microsoft in a threat intelligence report.
— A new generation of cross-border data transfer initiatives, and not merely those between the U.S. and the EU, could mark a new era of data sharing that upholds privacy standards while facilitating global trade, suggests Kenneth Propp for the Atlantic Council.
— The use of human subjects to train AI-generated responses led to a significantly improved service (in terms of a chatbot) and allowed for faster tweaks in the underlying model to weed out potential harmful/inappropriate interactions, according to research from a group of AI engineers at Meta.
SUBSCRIBE to the POLITICO newsletter family: Brussels Playbook | London Playbook | London Playbook PM | Playbook Paris | POLITICO Confidential | Sunday Crunch | EU Influence | London Influence | Digital Bridge | China Watcher | Berlin Bulletin | D.C. Playbook | D.C. Influence | Global Insider | All our POLITICO Pro policy morning newsletters
Appello da parte di Abune Tesfaselassie Medhin, Vescovo dell’Eparchia di Adigrat per scongiurare il blocco sugli aiuti alimentari in Tigray, Etiopia
Ambulanze in attesa di partire verso il luogo del massacro a Togoga – Tigray giugno 2021
Abbiamo trovato decine di feriti sdraiati sul ciglio della strada che aspettavano impotenti le ambulanze. Massacro di Togoga Tigray giugno 2021
Alcuni sanguinavanao, altri erano in blocco respiratorio… Massacro di Togoga in giorno di mercato Tigray giugno 2021
Soccorso ai feriti e alle vittime del massacro di Togoga Tigray giugno 2021
Abbiamo visto circa 50 persone ferite, tra cui una bambina di quattro anni. Togoga massacre Tigray giugno 2021
Direzione Togoga, giorno dopo il massacroin giorno di mercato in Tigray, giugno 2021
Quattro ambulanze sono finalmente arrivate alle 16:00 del 23 giugno 2021, molto tempo dopo: 27 ore dall’attacco aereo.
Irob woreda, Tigray Etiopia – giugno 2023
Alitena – ospedale bombadato dalle forze eritree
Piero Bosio
in reply to Informa Pirata • • •Ma perché un servizio importante come la telemedicina non lo gestisce il Servizio Sanitario Nazionale o Regionale?
Informa Pirata likes this.
Informa Pirata reshared this.