“We have two years to be able to hire the 80 people” when the AI Act is fully implemented. 

– an official on Wednesday about reorganising the European Commission’s DG CNECT. 

Story of the week: The European Commission’s Directorate General of Communications Networks, Content and Technology (DG CNECT) is reorganising its internal structure to accommodate the Artificial Intelligence (AI) Office, said officials on Wednesday. The AI Office is tasked with implementing the bloc’s landmark AI Act, which looks to regulate the use of the technology based on risk. DG CNECT’s A Unit, led by Lucilla Sioli, will morph into the AI Office, with some sub-units being re-assigned to AI-related tasks, effective 16 June. The reorganisation was already reported by Euractiv, along with questions on who will lead the office and how it will be staffed last week. “It’s not a huge reorientation,” as most of the units are already working on AI, said an official.  Read more. 

Don’t miss: A new compromise text of the draft law on online child sexual abuse material (CSAM), dated 28 May and seen by Euractiv, excludes audio communications from the scope and tries to strike a new balance between encryption and fighting CSAM. The new text completely excludes audio communications from foreseen detection orders, leaving visual content, images, videos, and URLs to be combed through and says CSAM should remain detectable across all interpersonal communication services, including those with end-to-end encryption. However, users must consent to this detection under the providers’ terms and conditions. France might support the new compromise, potentially eliminating the blocking minority and clearing the way for progress on the file within the Council, which Paris and Berlin have blocked, the people familiar with the file said. Read more   

Also this week: 

• Staffing questions swirl around Commision’s AI Office 
• Tech file on online child abuse material remains incomplete, risks gridlock 
• Picking up the AI Liability Directive after the tech policy spree 
• European Court of Auditors: Commission failed to effectively coordinate boost in AI investment 
• Panasonic patents trust files claim against Xiaomi in Paris 
• European startups call for a more integrated single market in declaration 
• Tech giants do not have to follow Italian authorities’ provisions, rules EU Court

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

Commission failures. The European Commission failed to effectively coordinate a planned boost in investments in artificial intelligence (AI) on par with global leaders between 2018 and 2022, said the European Court of Auditors (ECA) in a report published on Wednesday. The Commission declared AI “one of the most strategic technologies of the 21st century” in 2018 when it announced an initiative to boost investment in the technology and increase its uptake among the private and public sectors. The court said the difference between US and EU investment had doubled by 2020, citing a 2022 technical report by the Commission’s Joint Research Center. Read more. 

Picking up the AILD. The next European Parliament will have to determine whether a niche piece of liability rules related to AI, the Artificial Intelligence Liability Directive (AILD), can be useful after the previous parliament’s legislative spree. The responsible legal affairs committee has commissioned a study from the European Parliament’s Research Service, who in turn hired an external contractor to determine whether there is a legal gap left by the AI Act and PLD for the AILD to fill, Kai Zenner, digital policy adviser for MEP Axel Voss, who leads work on the AILD at the JURI committee, told Euractiv. At the Council level, there is uncertainty about the AILD’s usefulness, and Zenner said two member states have argued against the file. Read more. 

Drug giant Sanofi partners with OpenAI. Last Tuesday, French healthcare company Sanofi announced a partnership with OpenAI and Formation Bio. They aim to create an AI-driven software that speeds up the drug development process, making the delivery of new medicines to patients more streamlined and effective, reads the press release.   

The chatbot consultant. According to a Thursday announcement, PwC is set to become OpenAI’s largest customer and first reseller of ChatGPT Enterprise, a chatbot aimed at businesses. The tool will reach PwC’s 75,000 US and 26,000 UK employees under the agreement, terms for which were not disclosed.  

ChatGPT does not effectively support influence operations. OpenAI said the use of its tools in covert influence operators, including from China, Iran, and Russia, did not lead to increases in audience engagement or reach in its first threat report published on Thursday. The analysis comes while EU institutions and civil society are concerned about the use of generative AI to influence campaigns.  

Startups on the single market. Associations of European startups called for more integration of the EU’s single market to better support innovators in a political declaration on Thursday. Released seven days before the European Elections, June 6-9, underscores the importance of innovation for the bloc’s competitiveness on the global stage. Signatories include startup associations from Germany, France, Italy, Spain, Poland, The Netherlands, Bulgaria, Denmark, and Finland, as well as the global association Allied for Startups.  Read more. 

Operation against botnets. Operation Endgame, the largest ever operation against botnets, conducted between May 27 and 29, 2024, resulted in the takedown of over 100 servers worldwide and four arrests, Europol announced. The international effort targeted droppers such as IcedID, SystemBC, Pikabot, Smokeloader, and Bumblebee, which are used to deploy ransomware and other malware. The operation disrupted criminal infrastructures and froze illegal proceeds, significantly impacting the dropper malware ecosystem. Further actions and investigations are ongoing, with eight fugitives to be added to Europe’s Most Wanted list. 

Politician’s leaked info. Email addresses and sensitive information of 918 British MPs, MEPs, and French deputies and senators have been leaked on the dark web, said an investigation by Proton with Constella Intelligence, it was found that the e. British MPs were the most affected, with 68% of checked email addresses exposed, followed by nearly 44% of MEPs and 18% of French politicians. These breaches are not due to hacks of government systems but stem from politicians using their official emails on compromised third-party websites, a post by Proton, published on Thursday, reads. Exposed data includes plain text passwords, posing significant security risks. 

New cyber resilience centre. Mastercard has opened its European Cyber Resilience Centre at its European Headquarters in Belgium, the company announced on Thursday. This centre will serve as a hub for cybersecurity expertise, addressing the growing threat of cybercrime by bringing together national cyber intelligence centres, law enforcement, and industry bodies across Europe, the firm said.  

Temu’s VLOP designation. The European Commission designated online fashion retailer Temu as a “very large online platform” (VLOP) under the EU’s Digital Services Act (DSA) on Friday. The designation does not come as a surprise as the fast-fashion retailer already said it has 75 million monthly users per month in the EU, well above the DSA’s 45 million threshold. The online marketplace, operated by the Chinese e-commerce company PDD Holdings, has now been added to the DSA’s list of very large online platforms (VLOPs). However, as of 17 February, Temu, along with all online platforms in the EU, was already required to comply with the general obligations outlined in the DSA. Temu is “fully committed to adhering to the rules and regulations outlined by the DSA,” a company spokesperson told Euractiv. Following today’s designation, Temu will “have four months to comply with the most stringent obligations” of the DSA, Commission spokesperson Thomas Regnier said during the Commission’s regular midday briefing on Friday. This means that the deadline for Temu is the end of September. Read more.

Crackdown on Telegram? Telegram, the messaging platform increasingly used by purveyors of Russian disinformation, may face scrutiny under the European Union’s new crackdown on online platforms disseminating illegal and harmful content. Regulators are assessing whether Telegram should be classified as a major online platform under the EU’s new content moderation laws, according to sources familiar with the situation, Bloomberg reported on Tuesday. 

Italian semiconductors. The Commission approved on Friday €2 billion of Italian state aid to Geneva-based STMicroelectronics for the “construction and operation of an integrated chip manufacturing plant for Silicon Carbide power devices in Catania, Sicily.”  

Singapore’s data centre plans. The city-state announced on Thursday that it is investing S$300 million (€200 million) in quantum computing and developing a roadmap for sustainable data centres.  

Atos delays. Embattled French IT giant Atos delayed the shareholder meeting where its 2023 financial report to December 31 would have been approved, the firm said in a statement. The delay was approved by the Commercial Court of Pontoise “to provide Atos with a stable framework to complete the current discussions on a financial restructuring agreement by July 2024.”  

A look back at the CSAM file. The draft law on detecting and reporting online child sexual abuse material (CSAM), sparking criticism and tension in the past, remains one of the important yet incomplete tech files of the EU. Since February, member states have urged prioritising the implementation of existing digital policy over creating new bills – except the CSAM draft. Despite numerous compromise texts published by the Belgian Presidency of the Council of the EU, a general approach is unlikely to be reached by the end of its term in June. The Hungarian presidency will have to take over the file in July, but it is not expected to be a priority during its tenure. Read more. 

CrowdTangle and elections. On Monday, Meta announced adding an EU Parliamentary elections Live Display on CrowdTangle to view Facebook and Instagram posts for the upcoming elections.  

Pro-Russian Facebook ads. Research reveals that hundreds of pro-Russian ads are still circulating on Meta’s Facebook just before the European elections, despite an EU investigation, Politico reported on Thursday. The ads, part of a Russian influence campaign, have reached millions in France, Germany, Italy, and Poland, spreading anti-Ukraine and anti-EU narratives. Researchers criticise Meta’s inadequate response under the Digital Services Act (DSA) as the European Commission probes Meta for failing to curb disinformation.  

Meanwhile, in the US. Despite pressure from US lawmakers and regulatory scrutiny, TikTok is working on a distinct recommendation algorithm for its American users, aiming to untangle its US operations from its Chinese parent company, ByteDance, Reuters reported on Thursday. The endeavour involves separating millions of lines of code, which could take over a year to complete, with engineers tasked with ensuring independence from ByteDance’s broader network while eliminating any connections to Chinese users, the article reported. 

The French telecoms vision. In a speech on Thursday, the president of France’s regulator pushed back on some of the Commission’s plans for the telecoms sector, including deregulation, but she supported provisions on subsea cable resilience. Laure de la Raudière reiterated her call for a greener telecoms market, including more accountability for Big Tech’s environmental footprint.

Sun Patent Trust vs. Xiaomi. The 4G telecoms patents owner Sun Patent Trust has “summoned” Chinese smartphone maker Xiaomi to a Paris court, according to a Wednesday press release, likely looking for witness testimony from the European telecommunications standardisation body. The dispute revolves around Xiaomi’s allegedly unauthorised use of Long Term Evolution Advanced (LTE-A) technology to provide 4G communication in “nearly all 4G enabled devices it produced since late 2018.” Read more. 

Commission approves €22 billion KKR acquisition of Telecom Italia’s infrastructures. The Commission approved “unconditionally” under the EU’s antitrust Merger Regulation the acquisition by US investment company KKR of Netco, the fixed infrastructures of Telecom Italia. According to the Commission, the deal “would not significantly reduce the level of competition”. The deal should be finalised shortly. 

French telecom operator mulls acquisition of Luxembourg operator. French billionaire Xavier Niel’s investment vehicle Atlas Investissement, who owns Iliad, the parent company of French operator Free, is mulling buying Luxembourgish Milicom International, which operates in the US and nine Latin American states.  

Nokia wins. Nokia is set to win a contract to supply 5G radio equipment to Portuguese telecom operator MEO, according to an internal Nokia blog post and sources familiar with the matter, Reuters reported on Monday. This marks a significant shift as MEO has exclusively used Huawei for its 2G, 3G, and 4G Radio Access Network (RAN) equipment, Reuters reported.  

• China Creates $47.5 Billion Chip Fund to Back Nation’s Firms (Bloomberg) 
• France is an AI hub, but a wrinkle in tax policy is holding it back (Semafor) 
• How A.I. Made Mark Zuckerberg Popular Again in Silicon Valley (The New York Times) 

[Edited by Alice Taylor]

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