Digital standards

EU regulators seek common approach on DSA

The Coimisiún na Meán has warned that differing interpretations of the Digital Services Act (DSA) by EU regulators are hindering a unified approach to online platform regulation.

Maria Donde, Director of International Affairs at Coimisiún na Meán, highlighted the challenges of aligning various regulators’ approaches to the DSA, which has left room for interpretation.

She emphasised the importance of finding common ground, especially as the DSA, which came into effect last February, imposes transparency and election integrity requirements on platforms.

The DSA requires each EU member state to appoint a Digital Services Coordinator as a point of contact for platforms. Ireland, home to major platforms like TikTok and X, is at the forefront of enforcement.

Donde stressed the need for a consistent voice within the EU, particularly as the law faces criticism globally. The US government has condemned the EU’s regulatory approach, calling it a threat to free speech and accusing Europe of sidelining US tech companies.

The European Commission has already initiated several investigations under the DSA, targeting platforms such as X, TikTok, and Temu. These probes are ongoing, with potential fines for non-compliance reaching up to 6% of a company’s global turnover.

For more information on these topics, visit diplomacy.edu.

Fake movie trailer channels banned from monetization on YouTube

YouTube has taken action against AI-driven fake movie trailer channels, stripping them of their ability to monetize content. Following an investigation by Deadline, two of the most prominent channels, Screen Culture and KH Studio, have reportedly lost their ad revenue privileges.

With over two million subscribers and nearly two billion views combined, these channels created misleading trailers by splicing footage from existing films with AI-generated content.

Many unsuspecting viewers believed they were seeing genuine first looks at upcoming projects, such as Grand Theft Auto VI and Christopher Nolan’s The Odyssey.

Hollywood studios have reportedly lobbied YouTube to maintain monetization for such channels, though the reasons remain unclear. However, YouTube’s policies explicitly state that content must be ‘significantly changed’ and not copied solely for generating views.

While KH Studio’s founder defended their work as ‘creative exploration,’ Screen Culture’s founder questioned, ‘what’s the harm?’ YouTube’s latest crackdown suggests it is taking a firmer stance on AI-generated misleading content.

For more information on these topics, visit diplomacy.edu.

Elon Musk merges xAI and X to create XAI Holdings

Elon Musk’s AI startup xAI has officially acquired X, the social media platform (formerly Twitter), in an all-stock deal that values the two businesses combined at over $100 billion.

Musk disclosed that the transaction pegs xAI at $80 billion and X at $33 billion, though the latter includes $12 billion in debt, which brings its effective value to $45 billion.

The merged entity, XAI Holdings, streamlines Musk’s sprawling tech empire and solidifies the relationship between his AI pursuits and the platform that provides the training data.

According to Musk, the goal is to unify ‘data, models, compute, distribution and talent,’ enabling tighter integration between X’s reach and xAI’s growing capabilities.

This structural shift also clarifies to investors, many of whom have been concerned about X’s financial direction after Musk’s sweeping changes led to a loss of users and advertising partners.

Musk purchased Twitter in 2022 for $44 billion, which burdened the company with substantial debt. Since then, he has drastically altered the company’s operations and content policies under a ‘free speech absolutism philosophy,’ which has alienated many advertisers.

Although X’s advertising revenue dropped sharply post-acquisition, projections for 2025 show signs of recovery, with US ad sales expected to reach $1.31 billion, marking a 17.5% increase.

xAI, launched in 2023, has quickly positioned itself among leading AI labs. Its chatbot, Grok, has been trained using data from X, offering a competitive edge against other AI giants like OpenAI and Anthropic.

Analysts suggest that owning X gives xAI exclusive access to a rich proprietary data stream, something competitors lack. This advantage could strongly boost Grok’s development and positioning in the market.

Some investors in xAI, such as Andreessen Horowitz, Sequoia Capital, Fidelity, and BlackRock, also have stakes in X, making the merger a logical, if unexpected, evolution.

Financially, it also marks a turning point: banks that held onto Musk’s Twitter debt could finally sell it this year without losses, while X recently raised nearly $1 billion in new equity at a valuation close to its 2022 purchase price.

The merger may also influence broader industry trends. Analysts believe this move could inspire smaller social media platforms to seek strategic alliances with AI developers, especially given xAI’s high valuation.

One thing is clear: the XAI Holdings formation underscores a growing convergence between digital communication infrastructure and AI.

For more information on these topics, visit diplomacy.edu.

French watchdog fines Apple for abuse of app tracking tool

Apple has been fined €150 million ($162.42 million) by French antitrust regulators for allegedly abusing its dominant position in mobile app advertising between 2021 and 2023. The fine is the first to be imposed on Apple over its App Tracking Transparency (ATT) tool.

While the tool, which allows iPhone and iPad users to control app tracking, is not criticised itself, the French competition watchdog claimed its implementation was excessive and not proportional to its goal of protecting personal data.

The French regulators stated that ATT particularly harmed smaller publishers, who rely heavily on third-party data for their business. Despite the fine, Apple was not required to modify the ATT tool.

The decision follows complaints from online advertisers, publishers, and internet networks, who accused Apple of misusing its market power. Apple expressed disappointment with the fine but noted that no changes to the tool were mandated.

The fine comes after a €1.8 billion penalty last year from the EU, which accused Apple of restricting music streaming competitors. Additionally, the German antitrust agency has launched a probe into Apple for allegedly giving itself preferential treatment with the same privacy tool.

For more information on these topics, visit diplomacy.edu.

TikTok ban threatens 170 million American users

The US is just days away from imposing a ban on TikTok unless a deal is struck with its Chinese parent company ByteDance. The ban, set to take effect on Saturday, would affect 170 million American users of the popular app.

However, President Donald Trump has expressed confidence that an agreement will be reached in time. He extended the deadline from January to April 5 to give ByteDance more time to find a non-Chinese buyer for TikTok’s US operations.

Trump mentioned that there is significant interest from potential buyers, with private equity firm Blackstone reportedly evaluating a minority investment in TikTok’s US business.

The discussions are centred on ByteDance’s existing non-Chinese shareholders, including Susquehanna International Group and General Atlantic. Washington’s main concern is that TikTok’s ownership by ByteDance allows the Chinese government to potentially influence the app and collect data on Americans.

Despite the pressure, TikTok has yet to comment on the situation. If no agreement is reached by the deadline, TikTok faces the risk of being banned, though the app would remain on users’ devices if already installed. However, new users would not be able to download it.

The app is already banned in countries like India over similar national security concerns.

For more information on these topics, visit diplomacy.edu.

OpenAI faces copyright debate over Ghibli-style images

Studio Ghibli-style artwork has gone viral on social media, with users flocking to ChatGPT’s feature to create or transform images into Japanese anime-inspired versions. Celebrities have also joined the trend, posting Ghibli-style photos of themselves.

However, what began as a fun trend has sparked concerns over copyright infringement and the ethics of AI recreating the work of established artists instead of respecting their intellectual property.

While OpenAI has allowed premium users to create Ghibli-style images, users without subscriptions can still make up to three images for free.

The rise of this feature has led to debates over whether these AI-generated images violate copyright laws, particularly as the style is closely associated with renowned animator Hayao Miyazaki.

Intellectual property lawyer Even Brown clarified that the style itself isn’t explicitly protected, but he raised concerns that OpenAI’s AI may have been trained on Ghibli’s previous works instead of using independent sources, which could present potential copyright issues.

OpenAI has responded by taking a more conservative approach with its tools, introducing a refusal feature when users attempt to generate images in the style of living artists instead of allowing such images.

Despite this, the controversy continues, as artists like Karla Ortiz are suing other AI generators for copyright infringement. Ortiz has criticised OpenAI for not valuing the work and livelihoods of artists, calling the Ghibli trend a clear example of such disregard.

For more information on these topics, visit diplomacy.edu.

US fabs to catch up with Taiwan tech

TSMC says future chip factories in the US will take two years or less to complete, a big step forward from the five years needed for its first Arizona plant. The goal is to narrow the technology gap with its cutting-edge Taiwanese fabs.

While the first US fab makes chips on a 4nm process, TSMC aims to start 3nm production in 2028 and reach 2nm ‘before 2030.’ This would bring American output closer to the most advanced nodes used in Taiwan.

For Apple, which relies heavily on TSMC, the move reduces geopolitical risks tied to China–Taiwan tensions. Critics, however, point out that all R&D remains in Taiwan, limiting the US’s chances of true semiconductor leadership.

For more information on these topics, visit diplomacy.edu.

Commission seeks simpler, harmonised telecom rules

EU Tech Commissioner Henna Virkkunen has voiced support for using a Regulation, rather than a Directive, in the upcoming Digital Networks Act.

She says this would ensure consistent implementation across all member states, avoiding the patchwork seen under current telecom rules.

Virkkunen also hinted at easing merger rules and reducing ex-ante regulation within the existing framework, the European Electronic Communications Code.

These changes, she noted, could encourage investment and help the EU meet its goal of full 5G and fibre coverage by 2030.

She criticised slow national efforts to phase out high-risk Chinese components from 5G networks, calling for stronger action.

Her stance follows pressure from MEPs concerned about ongoing cybersecurity risks and lack of enforcement.

For more information on these topics, visit diplomacy.edu.

Korea delays decision on DeepSeek

Korean authorities say no decision has been made on when China’s DeepSeek AI app can resume operations in the country. The app was suspended last month due to concerns over its data handling practices.

Talks are ongoing between the Personal Information Protection Commission in Korea and DeepSeek, which recently appointed a local representative and pledged to comply with Korean privacy law.

DeepSeek is considered a key player in the Korean market, but officials stress that any resumption will depend on satisfactory privacy safeguards. No timeline has been set for lifting the suspension.

For more information on these topics, visit diplomacy.edu.

EU softens AI copyright rules

The latest draft of the EU AI Act’s Code of Practice offers a more flexible approach to copyright rules, focusing on proportionate compliance based on a provider’s size and capabilities.

However, this change comes as model providers face looming deadlines under the Act.

AI Developers must still avoid training on pirated content, respect opt-outs like robots.txt, and make reasonable efforts to prevent models from repeating copyrighted material.

However, they are no longer expected to perform exhaustive copyright checks on every dataset.

With potential fines of up to 15 million euros or 3% of global turnover, stakes remain high. Still, stakeholders welcome the clearer, more practical path to compliance, with final feedback on the draft due by the end of this month.

For more information on these topics, visit diplomacy.edu.