Welcome to Euractiv’s Tech Brief, your weekly update on all things digital in the EU. You can subscribe to the newsletter here.  

“The Presidency aims at reaching a partial general approach at the meeting of the Council on 10 October 2024.”

• Reads a document by the Hungarian presidency about the draft law to combat online child sexual abuse material (CSAM).

Story of the week: Resistance to the Hungarian presidency’s approach to the EU’s draft law to combat online CSAM was still palpable during a member states’ meeting on Wednesday. Poland, Germany, and other EU countries opposed the proposal, with the Netherlands withdrawing support altogether. Hungary aims to secure a general approach by October, with a detailed proposal expected soon, while the European Commission is set to provide its opinion to the presidency by 30 September. Read more.

Don’t miss: News over the the European Commission's internal search for a lead scientific adviser on Artificial Intelligence (AI) was met with some controversy this week. The Commission denied it ever said the position would be filled by an external candidate. Read more.

Also this week:

• Bruegel: The European Commission should build muscle to regulate killer acquisitions
• Dutch data protection watchdog hits Clearview AI with €30.5 million fine for misusing facial recognition data
• Civil society organisations call for EU-wide spyware ban
• Irish data privacy watchdog case against X is ‘struck out’
• EU, UK, US, Israel sign world’s first AI Treaty

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

World’s first AI Treaty. The EU, UK, US, and Israel signed the world’s first treaty protecting human rights in AI at a ceremony in Lithuania on Thursday. Civil society groups say the text has been watered down. After years of negotiations, the private sector is largely excluded from the Treaty, leaving mostly the public sector and its contractors under its scope. Read more.

NVIDIA bubble trouble. On Tuesday, NVIDIA stocks fell 9.5%, losing $279 billion in value – the largest single-day loss in value for a US stock. This came after the firm announced on 28 August a record quarterly revenue of $30 billion, more than doubling last year's figures but still falling short of expectations. The company also denied a Bloomberg report that the US Department of Justice had subpoenaed the company over antitrust violations.

Breton makes it to the 'Time 100 Most Influential People in AI 2024’. The Internal Market Commissioner who negotiated EU’s landmark AI Act, Thierry Breton, is the only EU politician featured in the Time's list, published on Thursday. California State Senator Scott Wiener, who authored the SB 1047 bill regulating AI, also made it to the list, along with Indian Digital Minister Ashwini Vaishnaw and New Mexico’s Martin Heinrich, a key member of the US Senate AI working group.

Safe Superintelligence. Previous Chief Scientist of OpenAI Illia Sutskever has raised $1 billion in funding for his three-month-old 10-employee startup Safe Superintelligence (SSI) according to Reuters. Sources close to the matter told Reuters the startup was valued at  was a $5 billion. Investors include Andreessen Horowitz, Sequoia Capital, DST Global, SV Angel and NFDG.

CERN for AI. Brussels-based think tank International Centre for Future Generations (ICFG) published a new CERN for AI proposal, including €30billion- €35 billion compute infrastructure investments on Wednesday. European Commission President Ursula von der Leyen endorsed the plan in the summer. Among several previously floated proposals, both proponents and critics of such initiatives requested more details. A CERN for AI “might be the EU’s last chance to catch up to foreign advanced AI developers,” the ICFG report said.

Meta and Brazilians’ personal data. Meta informed Brazilians how it will use their personal data to train generative AI, following a demand from Brazil's National Data Protection Authority (ANPD), Reuters reported on Tuesday. Starting Tuesday, users in Brazil received notifications via email, Facebook, and Instagram, allowing them to reject this use of their data. In July, ANPD suspended Meta’s privacy policy over AI data use but lifted the suspension after Meta agreed to issue disclosures. Meta's use of personal data to train AI models has sparked similar privacy concerns raised in the EU.

Cocktail party problem solved. Wave Sciences, a company using AI to separate overlapping sounds for forensics, military, and consumer tech, says it solved the “cocktail party problem” by developing AI that isolates individual voices in noisy recordings, the BBC reported on Wednesday.

Clamouring for the Code. The European Commission received interest from nearly 1,000 organisations and individuals looking to participate in drafting the first Code of Practice for General-Purpose AI models, like ChatGPT, the EU executive said on Friday. The AI Office is currently reviewing applications, with the process to be kicked off in a plenary on 30 September.

The debate around killer acquisitions is back. A day after the EU’s top court issued a ruling seemingly limiting the Commission’s authority to halt ‘killer acquisitions’  influential think tank Bruegel called for a stronger regulatory strategy from the EU’s competition watchdog. In a memo to the next competition commissioner, Bruegel said: “The Commission must ensure there is a strategy to review small acquisitions that nonetheless have a large impact on innovation and competition.” Read more.

Did Google breach competition rules, again? European Court of Justice Advocate General Laila Medina said that Google might be violating EU competition rules by refusing to let Enel X's electric vehicle charging app, JuicePass, work with Android Auto, according to a Court document dated Thursday.

Dutch DPA baring its teeth. The Dutch Data Protection Authority (DPA) fined Clearview AI €30.5 million on Tuesday for illegally building a database with over 30 billion photos. The authority warned against using the company's services, saying the use is prohibited. Similarly, the Dutch DPA fined Uber lat week for transferring European taxi drivers’ personal data to the US without proper protection for over two years, violating the GDPR. Read more.

Civil society wants a spyware ban. Organisations demanded comprehensive legislation banning spyware throughout the EU, citing widespread misuse and insufficient regulation, in a joint statement on Tuesday. Civil society groups said that existing regulations on press freedom are either insufficient or contain loopholes. Read more.

X proceedings are officially over. Court proceedings by the Irish Data Protection Commission (DPC) against social media platform X are officially over as of Wednesday after the company agreed to permanently stop processing some personal data collected in the EU for training AI. This follows a complaint filed by consumer organisations and the Irish DPC to file a case in Ireland’s High Court for alleged violations of the GDPR by X. Read more.

Problem solved? In a LinkedIn post, Marco Scialdone, a lawyer and adjunct professor, questioned whether X’s commitment to delete personal data used for Grok is sufficient. He highlighted concerns about the initial legality of data use and the persistence of data influence in AI models. Scialdone, who filed a GDPR complaint before the DPC court proceedings against X on behalf of consumer organisations, advocated for more comprehensive measures like algorithmic disgorgement.

YouTube and teen well-being. YouTube is working to improve teen well-being with new content limits to reduce exposure to harmful topics like body image and social aggression, according to the platform’s blog post from Thursday. A new supervised experience will link parent and teen accounts, offering insights and notifications on activity. The news follows ADWEEK’s August report claiming that Google sales representativess allegedly advised targeting teens with ads, potentially violating policies. The Financial Times also reported that both Google and Meta have been targeting 13- to 17-year-olds. The push for legislation to curb addictive design in Big Tech platforms will likely continue in the EU as regulators seek to address concerns about the impact of digital platforms on youth well-being.

X turnover. After major layoffs following Elon Musk's 2022 acquisition of X, the company is now hiring for its safety and cybersecurity teams. X's Global Head of Privacy at X, Renato Leite Monteiro, announced on LinkedIn this week that he is leaving his position after almost five years. X's global affairs head, Nick Pickles, also announced on Thursday that he is leaving his position after over ten years.

Snapchat letter. In a letter to employees on Tuesday, Snap CEO Evan Spiegel acknowledged concerns about slower advertising growth compared to competitors but emphasised the shift towards a more stable and diversified ad business. Spiegel wrote about new initiatives, including innovative ad placements and advancements in machine learning and generative AI. He also discussed maintaining a conservative financial approach amid global economic uncertainty.

Bluesky-rockets to the top. Bluesky gained 2.4 million users after X's shutdown in Brazil, topping app charts, TechCrunch reported on Tuesday. Downloads soared globally, including in the EU, where countries like Portugal and Spain saw sharp increases.

Microsoft takes steps against deepfake porn. The company is partnering with platform StopNCII to help users scrub AI-generated sexually explicit images of themselves, according to a Thursday blog post. Microsoft has already taken action against 268,899 images, it said.

ETNO DG steps down. Lise Fuhr, director general of the European Telecommunications Network Operators' Association (ETNO), Europe's leading telecom association of the largest telcos, will step down in the autumn. Fuhr will take over as CEO of GÉANT, an organisation supporting research and education networks. ETNO is expected to reveal new leadership plans later this month, along with their future plans, the association said on Tuesday.

French telecom tycoon is Ukraine’s top foreign investor. As reported by Challenges on 30 August, the Ukrainian telecom regulator (Antimonopoly Committee of Ukraine, AMCU), on 25 July, allowed Xavier Niel’s DVL Telecom to acquire Ukrainian third mobile operator Lifecell, mobile company Global Beeline, and tower company Ukrtower. Earlier in 2024, the Ukrainian regulator allowed DVL Telecom to acquire the Ukrainian national and largest fixed-line provider Datagroup Holding Limited. DVL is expected to invest over €1 billion, making it the top foreign investor in Ukraine, wrote Challenges.

• Antony Blinken Dragged US Diplomacy Into the 21st Century. Even He’s Surprised by the Results (WIRED)
• Spyware vendors thwart restrictions with new names and locations (Washington Post)
• AI’s impact on elections is being overblown (MIT Technology Review)

[Edited by Martina Monti]

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