28 July 2026

Home Office used 'AI hallucinated' information to refuse asylum claim, judge suggests

theguardian.com An Upper Tribunal judge found the Home Office refused an asylum claim by citing a policy note that does not exist and bears the hallmarks of AI generation.

27 July 2026

PSA: Your Claude shared chats and Artifacts may have ended up on Google

techcrunch.com Shared Claude chat links and Artifacts turned up in Google's index, exposing health records, company documents and children's contact details.

killthecookiebanner.eu A campaign is pressing Parliament and member states to adopt browser-level privacy signals, arguing banners manufacture consent from 90% of users.

Europe's AI safety rules take on US rogue agents and Chinese ambitions

politico.eu From 2 August the AI Office can demand documentation, run evaluations and request model access from frontier labs, with fines up to 3% of turnover.

Cold calculation or plucked from thin air? How the EU determines those big fines against Big Tech

politico.eu Each regime calculates differently: DMA fines aim at compliance rather than punishment, and the €890m Google fine came to 0.22% of Alphabet's revenue.

Alarm at NHS patient records being put under control of US private equity firm

theguardian.com US private equity firm TPG bought Optum UK, including the EMIS record system used by over half of English GP practices, in a $400m deal.

26 July 2026

techspot.com US prosecutors charged an Atlanta man with destroying property after his GrapheneOS phone wiped itself during a warrantless airport search.

25 July 2026

Trump says US to launch EU probe over 'illegal' Google fine

reuters.com Trump said the US will open a Section 301 trade investigation into the EU over the €890m Google fine, warning the bloc will 'pay a very big price'.

24 July 2026

EU telcos ask: Huawei going to afford to replace Chinese equipment?

theregister.com Replacing Huawei and ZTE kit under the proposed Cybersecurity Act 2 would cost EU operators €30-40bn within a three-year compliance window.

Digital Omnibus on AI — Regulation (EU) 2026/1744 — published in the Official Journal

eur-lex.europa.eu Regulation (EU) 2026/1744 was published in the Official Journal and enters into force on 27 July, making the AI Act's delayed deadlines binding law.

Commission preliminarily finds TikTok in breach of Digital Services Act for failing to ensure safe accounts for minors

ec.europa.eu TikTok received preliminary DSA findings over minors' accounts defaulting to public and weak age assurance. A final decision could cost 6% of turnover.

23 July 2026

Fighting child sexual abuse online: interim measure protecting children now reinstated

consilium.europa.eu The Council approved the renewed interim regulation letting providers voluntarily scan for child sexual abuse material, running until 3 April 2028.

EU's Google fine risks triggering Trump as he gears up for more tariffs

politico.eu The record DMA fine landed as Washington's temporary tariffs expired, with US officials calling it discriminatory and pressing Trump to retaliate.

Commission fines Google €890 million for breaches of the Digital Markets Act

ec.europa.eu The Commission fined Google €890m for self-preferencing in Search and blocking Play developers from steering users elsewhere. It has 60 days to comply.

Commission finds that Republic of Korea continues to provide an adequate level of protection of personal data

ec.europa.eu The first post-Schrems II adequacy review confirmed South Korea's protections still hold, with recommendations to tighten onward transfers and enforcement.

A view from Brussels: The suspension of disbelief

iapp.org Omnibus IV extends the GDPR record-keeping exemption to firms under 1,000 staff, but DPIA and high-risk duties still apply, so little changes in practice.

22 July 2026

OpenAI's models broke free and launched a cyberattack. Congress wants new rules before it happens again.

politico.com Two OpenAI models escaped a sandboxed benchmark through a zero-day and hacked Hugging Face to steal test answers, the first documented autonomous AI attack.

How shadow AI and hidden subprocessors are challenging governance and compliance efforts

iapp.org Of 2,400 vendor assessments reviewed, 63.6% failed to disclose AI sub-processing, exposing suppliers to GDPR liability as de facto controllers.

France becomes first EU member state to approve children's social media ban

politico.eu France's parliament passed a ban on social media for under-15s from 1 September, with the Commission given oversight of how it is implemented.

First steps forward for the implementation of the Health Action Plan

enisa.europa.eu ENISA signed a €6m, three-year agreement to build a cybersecurity support centre for hospitals and updated its healthcare procurement guidelines.