Tech Brief: DSA enforcement, German data strategy

DSA enforcement, German data strategy

“We are working hard on this @ThierryBreton”
– wrote Elon Musk, owner of X, on 25 August, the day of the Digital Services Act enforcement.

Story of the week: While the European Commission’s directorate in charge of implementing the Digital Services Act (DSA) has yet to be finalised, the new team is already besieged by challenges, including legal complaints, NGOs’ demands and alleged technical limitations. The legal complaints were filed to the Court of Justice of the European Union in June by Zalando and in July by Amazon. A group of 56 civil society organisations issued a letter to the Commission regarding election manipulation on 23 August. Eventually, the Computer & Communications Industry Association sent a letter on 29 August, focusing on legal and technical issues of the current Transparency Database. Read more.

Don’t miss: Germany launched a new data strategy to untap the potential of data by using it more effectively, as the Ministry of Economy discovered that 80% of industrially generated data is not used further. “So far, data has far too often remained unused and thus lacked digital innovation. This applies to industrial as well as public data. We want and need to change that,” Digital Minister Volker Wissing emphasised. The new National Data Strategy builds on the existing Data Strategy 2021 and aims to strengthen digital innovation and improve German competitiveness.” The data strategy complements the requirements from Europe and now forms the second important guideline for data policy – with clear timelines and mandates for action for us in Parliament and the German government,” Green MP Tobias Bacherle told EURACTIV. Read more.

Also this week:

Before we start: If you just can’t get enough tech analysis, tune in on our weekly podcast.

Spain creates AI watchdog. A decree published after the Council of Ministers meeting on 22 August officialised the creation of its national authority supervising Artificial Intelligence (Aesia). The document states that the new AI watchdog aims to “provide a framework of reference for the development of Artificial Intelligence.” The authority will be affiliated with the Department of Digitalisation and Artificial Intelligence within the Spanish Ministry of Economic Affairs and Digital Transformation.

UK takes all its friends to the AI summit. China may be invited to the first global summit about AI, which is expected to focus on safety and will take place in London on 1 and 2 November. While the news is not officially confirmed, several news outlets reported that Rishi Sunak’s administration would like to involve the country despite the disapproval from the European Union, Japan, and the United States. Foreign Secretary of the United Kingdom James Cleverly also shared a video from Beijing on X, in which he said that his trip there was, among other things, about “making sure that we address collectively both the opportunities and the challenges of AI and other technologies.” Meanwhile, the British prime minister said that the UK must introduce new legislation on AI in November, or it risks falling behind the EU and the US.

AI-based facial recognition. In the United Kingdom, the Home Office mulls to increase police use of facial recognition technologies. Its ambition would be to have a functionally ready biometric system in one to one and a half years. The government expressed its interest in technologies that could be deployed on existing surveillance infrastructures, such as CCTV cameras, while being able to identify individuals through facial recognition. The debate is gaining momentum in the UK, where it is unclear whether there is a legal basis for using such technologies, which have already been adopted by schools, police departments or private owners. In addition, the current situation contrasts with the path the European Union is following in opening its third session of trilogues on the AI Act and moving towards much stricter rules regarding facial recognition software.

Is this real life, or is it just an AI picture? To prevent misinformation, developed by DeepMind, Google’s AI arm, a beta version of SynthID was launched on Tuesday to try to identify pictures created by artificial intelligence. SynthID will embed a watermark into images created by Imagen, Google’s text-to-image generator. These watermarks will be permanent but invisible to the human eye so that they cannot be, for example, simply cropped out of a photo. The tool will also be able to determine the likelihood of an image being created by Imagen. DeepMind said, however, that the technology is not “foolproof against extreme image manipulation.”

Is it art or is it AI? To drive awareness for the influence of AI on art and culutre and bridge their gab, Ars Electronica and CultTech Association, two Austrian organisations are in negotiations to open a research centre called “AI for culture” which will be based in Austria. “While many feel intimidated by the effect tech is having on our societies and what we create in our society, we at CultTech are looking to embrace the empowering effect tech is having in the arts and culture realm,” says Dmitry Aksenov, CultTech founder.

France vs. GPTBot. Following steps by many English-language media, many French media groups have decided to block OpenAI’s web crawler, GPTBot, from collecting their content online. This includes all the France Médias Monde websites, such as France24.com, RFI.fr, mc-doualiya.com, and Radio France and TF1. On 8 August, OpenAI announced that the tool will automatically collect data from the entire internet to train its GPT-4 and GPT-5 models, filtering out paywall-restricted sources, any source that violates OpenAI’s policies, or those that gather personally identifiable information. The collected material, however, may include copyrighted content. Read more.

Debate over AI copyright exemption. On 30 August, the Culture, Media and Sport Committee of the Parliament in the United Kingdom called on Rishi Sunak’s government to “abandon artificial intelligence copyright exemption to protect creative industries”. The Members of Parliament are concerned about allowing Artificial Intelligence bots to “mine private intellectual property for profit without sharing with the original creators”, including the music, literature and works of art industries.

US-music industry brokers AI deal. According to the Financial Times, Universal Music and Warner Music entered negotiations with Google in the United States to receive copyrights on AI-generated musical content. The music industry sees AI technologies that mimic artists’ voices as a big threat. Google and Universal Music’s goal would be to develop an AI tool that allows people to create tracks legitimately and pay copyright owners for it. For now, such notorious “deepfakes” songs mixed Frank Sinatra’s voice to sing “Gangsta’s Paradise”.

You can’t merge with us. As EURACTIV reported, in May, EU antitrust authorities approved Microsoft’s proposed acquisition of gaming company Activision Blizzard just weeks after regulators of the United Kingdom’s Competition and Markets Authority (CMA) blocked Microsoft’s biggest takeover. The decision followed a lengthy investigation into whether the merger would threaten competition in the gaming market. Last week, however, Microsoft announced that they “are restructuring the transaction to acquire a narrower set of rights,” including “executing an agreement effective at the closing of our merger that transfers the cloud streaming rights for all current and new Activision Blizzard PC and console games released over the next 15 years to Ubisoft Entertainment SA, a leading global game publisher.” The company added that “the rights will be in perpetuity.” However, the CMA did not change its mind about blocking the merger. EURACTIV contacted the European Commission about whether this would affect the EU antitrust authorities’ decision. A spokesperson said they are “carefully assessing whether the developments in the UK require another notification to the Commission”.

Microsoft Teams and Office divorce. Just a month after the European Commission opened an antitrust investigation into bundling tactics by Microsoft, the company decided on 31 August to sell its communication and work platform Teams and its computer software Office 365 separately starting in October. The European Commission still needed to comment on the current decision. Concretely, this means that organisations will be able to buy licences for Teams only if they want to. Francisco Mingorance, executive secretary at CISPE, a cloud trade association, reacted by saying that this decision was “just the tip of the iceberg”, calling for further changes in Microsoft cloud market practices. Eventually, he said that CISPE “has identified over a dozen of additional” practices from the US tech giant.

Meanwhile, in Hungary, cyber attacks. According to the International Press Insitute’s (IPI) article, published on Tuesday, more than 40 Hungarian media websites hit by distributed denial-of-service attacks in the last five months, “a form of cyber-attack which temporarily slows or crashes websites by overloading their servers with millions of simultaneous access requests, leaving readers unable to access news and information for hours at a time.” The motive of these attacks is unknown, but IPI reports that most of the targeted sites include many of the country’s leading independent media, while no media outlet supportive of the ruling party has been targeted so far.

Should I stay, or should I block you now? After concerns emerged about the Russian ride-hailing service Yango (or Yandex) sending customer data to the Russian Federal Security Service, according to the European privacy authorities, the service will eventually be allowed for the time being to send data of its Finnish and Norwegian users to the Russian Federation. Yet, back in the middle of August, the Finnish and Norwegian regulators said they had banned the Russian service and its Netherlands-based partner Ridetech International from transferring data to Russia. Yango operates in over 20 countries, such as Israel, the United Arab Emirates, Algeria, Pakistan, Armenia, Georgia, Kazakhstan, Azerbaijan, Norway, and Finland.

5 VLOPs did not efficiently tackle Russian disinformation. According to a Commission report publicly available by mistake during less than 24 hours on Tuesday, Facebook, Instagram, Telegram, TikTok, X, and YouTube did not comply with the risk reduction rules provided in the Digital Services Act. These conclusions were drawn on actions undertaken between February and November 2022, at a period when the DSA did not legally apply to VLOPs. Since 25 August, all 19 designated “very-large online platforms” and “very-large search engines” are required to counter disinformation or risk paying hefty fines efficiently. On Wednesday, the Commission made the document available again.

France grumbles at Twitter “notes”. BFMTV has it that several French politics denounced the new contextual “notes” that X implemented under its posts. Ecologist Member of Parliament Sandrine Rousseau called it a “new type of harassment”. According to Member of Parliament Julien Bayou, the new X feature is still sensitive to astroturfing, a practice that hides sponsors of a message to make it appear as though grassroots participants supported it.

Fiona Scott Morton expresses her deception. In an article published by The Telegraph on 28 August, Morton, the US citizen who turned down a top Commission job, following a backlash spearheaded by the French government, said she was “surprised” and “disappointed” by these reactions. She called France “insecure” and doubted the European Commission’s independence. Contacted by EURACTIV, Digital Minister Jean-Noël Barrot, Economy Minister Bruno Le Maire, and President Emmanuel Macron, who openly criticised her nomination, did not wish to comment. Thierry Breton, Commissioner for the Internal Market, who, for some, used this controversy to his personal advantage and would now be eyeing the Presidency of the European Commission in 2024, did not comment to EURACTIV either.

Stalled start. Trilogues for the platform workers’ directive had hardly even started, and already they are stalling. The second round of interinstitutional negotiations, initially scheduled for 5 September, was pushed back to the 18th, as there was no sufficient progress from the last technical meeting, EURACTIV has learnt. EURACTIV can also confirm no interinstitutional text has been agreed to date.

Who cares about chips? Broadcom, a semiconductor manufacturer that provides hardware for infrastructure such as data centres and broadband networks and is also one of the world’s biggest chipmakers, signalled recently that the demand for chips remains stagnant. According to Bloomberg, since 2020, the projected gain would be the slowest now.

Commission’s new DSA team enforcement challenges. While the European Commission’s new directorate in charge of implementing the Digital Services Act has not yet been entirely completed, it already has to address a number of challenges, from cooperation with other entities to harmonisation within the bloc and potential skill shortages. Read more.

Meta shut down Chinese Spamouflage. 7,704 Facebook accounts, 954 pages, 15 groups, and 15 Instagram accounts were shut down by Meta recently. The accounts were associated with a Chinese political spam network that had targeted users in several parts of the world and usually posted positive commentary about China and Xinjiang province while criticising the US, Western foreign policies, journalists, and researchers. “This network originated in China and targeted many regions worldwide, including Taiwan, the United States, Australia, the United Kingdom, Japan and global Chinese-speaking audiences,” the tech giant said. In addition to the company’s social media sites, more than 50 other platforms were targeted as well, such as YouTube, TikTok, Reddit, Pinterest, Medium, Blogspot, Livejournal, X, VKontakte, Vimeo, and other smaller sites.

But there’s more: Meta also blocked Russian website domains. In addition, Meta also blocked thousands of website domains and attempts to run fake accounts and pages connected to the Russian operation, Doppelganger. Doppelganger “focused on mimicking websites of mainstream news outlets and government entities to post fake articles aimed at weakening support for Ukraine,” the company said. While it initially targeted France, Germany, and Ukraine, it also included the US and Israel. According to Meta, “this is the largest and the most aggressively-persistent Russian-origin operation we’ve taken down since 2017,” mentioning that they first disrupted Doppelganger a year ago.

TikTok removed a Chinese disinformation group as well. Following Meta’s steps, and after Guardian Australia reported the accounts to TikTok, the Chinese-owned platform also removed 284 accounts associated with a Chinese disinformation network for violating company policy against covert influence operations.

Trade associations are not impressed. Coordinated by CCIA Europe and DOT Europe, a group of nine trade associations published a joint statement on the European Media Freedom Act. According to the paper, the must-carry obligation risks allowing harmful content to stay on a platform for 24 hours and proliferate, contrary to VLOPs’ (very large online platforms) obligations to mitigate systemic risks under the Digital Services Act. They also find the overlap with DSA provisions problematic when challenging VLOPs’ content moderation decisions. Moreover, they find the platform-to-businesses and the DSA complaint-handling processes unnecessary. The statement urges MEPs to consider the negative impact of such provisions.

End of the road for Paris e-scooters. Starting 1 September, all 15,000 self-service electric scooters are banned from Paris. The city hall calls the move revolutionary amid an urban jungle, while stakeholders remain divided on the pros and cons. Read more.

Meanwhile in the Parliament. Several informal documents about the draft law aiming to prevent child sexual abuse material, seen by EURACTIV and distributed during this week’s technical meetings in the Parliament, put focus on the detection orders, voluntary and mandatory measures, as well as software application stores. Read more.

AI-assisted surveillance for Paris Olympic Games. The French decree putting in place the technicalities for algorithmic treatments of pictures taken by drones or CCTV was published on 28 August. The AI systems used during the Olympic Games should not use biometric data, yet be able to detect suspicious behaviours, such as hazardous crowd movement, presence of weapons, abandoned objects, people lying on the ground or the start of a fire. The estimated bid has been estimated at €2 million.

French Senate puts pressure on TikTok. Adam Presser, who previously served as chief of staff, was appointed as TikTok’s new head of operations in June and reports directly to TikTok’s global CEO, Shou Zi Chew. However, l’Informé said this week that he was appointed president of TikTok SAS, the French branch of Bytedance, following pressure from the French Senate. As EURACTIV reported in June, Claude Malhuret, the rapporteur for the Senate’s Commission of Inquiry, said that TikTok was registered in the French commercial register as a simplified company with a single shareholder, i.e. a company 100% lawfully owned by the entrepreneur Zhao Tian, who is also vice president of the Chinese application Toutiao, a subsidiary of Douyin, the Chinese version of TikTok. At the time, Eric Garandeau, TikTok’s public affairs director in France and Marlène Masure, TikTok’s Managing Director of Operations for France, Benelux and Southern Europe, said that neither had “ever met Ms Zhao Tian”.

X to allow political ads again. On Tuesday, the social media platform X announced in a blog post that they are hiring “to focus on combating manipulation, surfacing inauthentic accounts and closely monitoring the platform for emerging threats” ahead of next year’s US presidential election. The platform, still known at the time as Twitter, disallowed political ads in 2019. The blog post also added that they will focus “on combating manipulation, surfacing inauthentic accounts and closely monitoring the platform for emerging threats” while updating their policies.

Don’t copy me. The Danish Media Association filed a lawsuit against Google on Thursday. According to the association, the tech giant violated copyright and marketing laws by copying and making job ads from Jobindex, Denmark’s largest job portal, available on Google for Jobs without permission last year. Jobindex has failed to get Google to acknowledge this, so the job portal is now taking action. For the Danish court, this is the first time that a case involving the rules in the EU Copyright Directive Article 17 about platforms’ liability for content uploaded to their services will be tried.

Netflix under tax evasion indictment. The French subsidy of Netflix has been under tax fraud investigation since 2022. The main accusation is that the streaming giant used a financial arrangement between 2019 and 2021 that would have allowed it to “minimise its taxation by declaring in the Netherlands its French generated revenue”. According to La Lettre A, which revealed the tax audit on 30 August, civil servants are trying to verify if Netflix did not continue its illegal practices by comparing a dubious 2% margin in France against 20% in the United States. “We comply with the tax rules of all the countries in which we operate around the world,” a Netflix spokesperson told the AFP later during the day.

An Amazon-Shopify deal. Announced on Wednesday, Amazon and Shopify, two companies selling e-commerce software and services to brands and merchants, have struck a deal that will allow Shopify merchants, who are paying for Shopify’s e-commerce tools, to use Amazon’s logistics network. Previously, those merchants who used ‘Buy with Prime’, Amazon’s checkout service, were warned that this prevented the company from protecting against fraud and violated Shopify’s terms of service. Now, the payments will be processed through this checkout service.

Fair share momentum. Telefónica published a blog post on 28 August, defending the idea of a fair share tax on large internet traffic generators. Three days later, on 31 August, the European Telecommunications Network Operators’ Association co-signed a letter with its Korean counterpart (KTOA), stating that “large traffic generators should make a fair contribution to the sustainability and development of the connectivity networks”.

Saudi Arabia invests in European towers. On 28 August, United Group BV, a Dutch telecommunication company, completed the sale of its 4,800 towers across Bulgaria, Croatia and Slovenia to TAWAL, a Saudi TowerCo, for €1.2 billion. TAWAL, a large infrastructure provider in the Middle East region, is expanding its operations to the EU telecom market.

What do citizens across the EU expect? Among the top four priorities, European citizens value secure and sustainable digital infrastructure, reveals Cisco. According to its 2023 Broadband Survey, three quarters of 16,625 survey participants from the UK, Germany, Italy, France, Switzerland, Poland, Spain, Sweden, and the Netherlands consider broadband a critical national infrastructure and value sustainable and secure internet for all. Particularly, the carbon cost of connectivity is a concern for the majority. They are willing to pay more for sustainable broadband, reveals the survey.

Alina Clasen and Théo Bourgery-Gonse contributed to the reporting.

Large language models aren’t people. Let’s stop testing them as if they were. (MIT Technology Review)

“Have all potential loopholes for money laundering with digital payment services been properly assessed?” asks Charles Cuvelliez, Professor at the Belgian ULB University, in an op-ed in Le Monde.

Dans les groupes WhatsApp des ministères : « Je vous préviens, ça va être déceptif » (l’Insitut National de l’Audiovisuel)

[Edited by Alice Taylor]

***

Thanks for reading. Be sure to spread the word and come and say hello on Twitter.

***

Copyright © 2023 EURACTIV Media Network BV, All rights reserved.
You are receiving this email because you subscribed to receive email newsletters from EURACTIV.
Our mailing address is: