Countries and companies looking for global AI governance as US reaffirms opposition

Chinese AI labs ‘targeting products over frontier dominance’

The Financial Times notes that the week-long holiday for the lunar new year has seen the release of a series of new AI models as Chinese labs seek a competitive edge over their US rivals. Importantly, they are focused on models that are useful for building applications, according to Ritwik Gupta, an AI researcher at the University of California, Berkeley. This contrasts with US labs which view it as a ‘race for frontier dominance first, product second’, he said.  Observers cite ByteDance’s new video-generating model, Seedance 2.0, which has spooked Hollywood with its ability to produce multi-scene clips with realistic imagery and synchronised sound effects. Alibaba has released a model targeted at developers of AI agents. The Chinese companies are exploiting opportunities by offering greater freedom than US companies such as Anthropic and OpenAI, which place restrictions, such as usage limits, on developers. Additional note: In a post on X, Anthropic claims to have identified ‘industrial-scale distillation attacks on our models by (Chinese companies) DeepSeek, MoonshotAI and MiniMax. These labs created over 24,000 fraudulent accounts and generated over 16 million exchanges with Claude, extracting its capabilities to train and improve their own models’.

Anthropic sues US government for calling it a risk

AI firm Anthropic has taken the unprecedented step of taking legal action after it was described by the US government as a ‘supply chain risk’. The US Department of Defense has been in dispute with Anthropic after the company refused to allow the military unfettered use of its AI models. Anthropic has argued that the constitution ‘does not allow the government to wield its enormous power to punish a company for its protected speech’, and that ‘No federal statute authorizes the actions taken here.’ A White House spokeswoman said that ‘under the Trump administration, our military will obey the United States Constitution – not any woke AI company’s terms of service.’

North Korea using AI to deploy ‘fake workers’

Cyber experts describe North Korea as global experts in the ‘fake worker’ phenomenon, using AI to deceive firms into hiring North Korean operatives. The scam typically involves stealing an identity such as a dormant Linked In account or paying account holders for access. Generative AI is then used to create digital avatars and make deepfake video filters to appear in remote job interviews. North Korean fake workers have infiltrated more than 300 US firms, according to the Department of Justice, and Amazon recently announced that it has prevented more than 1800 suspected North Korean operatives from getting jobs since April 2024. Threat intelligence experts say that the phenomenon is increasingly spreading to Europe.

EU to mandate exclusion of ‘high-risk’ suppliers

Chinese-made equipment is set to be phased out from use in critical infrastructure under the EU’s new cybersecurity proposal. The move would bar Huawei and ZTE from telecoms networks as well as solar energy systems and security scanners. The move would revise the existing voluntary regime. A previous draft of the Cybersecurity Act noted that ‘fragmented national solutions have proven insufficient to achieve market-wide trust and co-ordination’. Industry representatives have expressed concern about the lack of viable alternatives – 90 per cent of solar panels installed in the EU are made in China for example, while telecoms operators have warned about the impact on consumer prices

European Commission’s ‘credibility on the line’ over Google investigation

European publishers and tech firms have written to EU leaders urging them to complete their two-year investigation into Google and issue a fine. The case concerns Google’s alleged favouring of its own services in online searches in breach of the Digital Markes Act, and was opened in March 2024. EU regulators says that they aim to complete DMA investigations within 12 months, but the case is thought to have been complicated by clashes with the US over attempts to curb the influence of big tech. The letter, from the European Publishers Council, calls for the investigation to be finished in the next week, arguing that the credibility of the European Commission is ‘on the line’, and that, ‘it is important that sustained pressure to dilute the DMA is not shown to have succeeded’.

Australian social media ban ‘is not improving teen safety’

Meta has deleted over 500,000 Australian accounts in order to comply with the country’s under 16s ban. The company said that it removed 330,000 users from Instagram and 173,000 from Facebook between 4 and 11 December. In its statement Meta argued that early data suggests the ban is not improving teen safety or wellbeing. Age verification measures are inconsistent and pushing teenagers towards smaller, less regulated apps, the company said.

X under continued pressure over Grok’s sexualised image generation

Governments around the world have condemned the sexually explicit content generated by Grok on X, including ‘undressed’ depictions of minors. Grok has restricted image generation and editing to paying subscribers, allowing users behaving illegally to be easily identified. However, regulators are questioning whether the company has gone far enough. In the UK, Ofcom has announced an investigation into whether Grok has complied with obligations under the Online Safety Act, and French regulator ARCOM is similarly looking at the company’s compliance under the Digital Services Act. Germany’s media minister has urged the EU to take legal action to stop the ‘industrialisation of sexual harassment’ on X. Action is also being taken in India, Malaysia and Australia.

In brief

President of the European Council António Costa has indicated that EU is open to the idea of more telecoms consolidation in Europe. Costa said that he had heard agreement that ‘in some sectors like telecom, we should allow for a degree of company consolidation to achieve the necessary levels of investment and innovation. This should be part of a kind of social contract’. Revised EU merger guidelines are expected to be published in April.

The Federal Communications Commission is imposing new national security reporting requirements across the US telecommunications sector. The aim is to address the risk of foreign control of FCC-issued licences. Any entity under ‘FAC’ (foreign adversary control) will have to make specific disclosures to the FCC and be held to ongoing reporting obligations.

AI will be a bubble unless benefits spread beyond tech companies, according to Microsoft Chief Executive Satya Nadella. He warned that the long term success of artificial intelligence technology depends on it being used by a broad range of industries and in the developed world. Speaking at the World Economic Forum in Davos, he said that ‘a tell-tale sign of if it’s a bubble’ would be if only tech groups were benefitting from the rise of AI.

OpenAI has announced that it will begin showing ads in the free and lower-priced tiers of its ChatGPT platform for some users in the US as the company seeks to increase revenues to fund its datacentres . The company says the advertising won’t affect any of ChatGPT’s answers, nor will data be shared with marketers.

Setting up the digital euro will cost European banks between 4 and 6 billion euros over four years, according to a European Central Bank governing council member. Piero Cipollone also said that the ECB would incur set-up costs of 1.3 billion euros and operational costs of c.300 million euros per year. The digital euro is seen as a way of keeping public money relevant in a digital economy and curb Europe’s fragmentary payments landscape, as well as restrict the role of non-EU providers.

The UK government has announced new laws as part of a wider crackdown on violence against women and girls. The plans impose a number of obligations on tech companies, including a requirement to take down abusive images from the internet within 48 hours and to detect and remove intimate images shared without content. Failure to act could result in services being blocked and fines of up to 10 per cent of worldwide revenue.

President Karol Nawrocki has vetoed legislation designed to enable the implementation of the Digital Services Act (DSA)  in Poland, despite the bill having been adopted by parliament at the end of 2025. Nawrocki says that the bill gives excessive powers to government officials to decide what content is allowed online, arguing that there is insufficient judicial oversight.

Sources: The Financial Times, Reuters, The Wall Street Journal, Euronews, Euractiv, US News, CNN,  TechCrunch, The Guardian, The Daily Telegraph, Bloomberg, Economic Times, BBC, Politico, Telecoms.com