United Kingdom
Google Cloud outage disrupts UK banks and businesses
- A power-related failure at a Google Cloud data centre in London disrupted services across the UK, affecting banks, businesses, delivery companies and court hearings. Google later confirmed that the issue had been resolved and services restored.
- Revolut, Lloyds Banking Group services, including Halifax and Bank of Scotland, and other platforms experienced issues. Some customers could not access funds or complete transactions, while HM Courts and Tribunals Service reported issues with its remote-hearing video platform.
- The outage also affected small businesses that rely on cloud services and connected platforms. The Veg Box Company could not process customer orders or arrange DPD parcel collections, resulting in lost sales and potential delivery delays.
- This highlights the widespread dependence on a small number of major cloud providers and how failures can impact essential services.
IC reprimands two Welsh organisations after ransomware attack on children data
- The Information Commission (IC) reprimanded Ogi Networks Limited and Data Cymru after a ransomware attack in November 2024 exposed information belonging to 3,300 people, mostly children. The attackers encrypted 182.7GB of data and extracted 42GB linked to projects supporting children and families.
- The breach exploited a test account that had remained available for more than 12 years and used the extremely weak password “Password1”, which could reportedly be cracked in less than a second. The account had not been removed despite no longer being needed.
- The IC found that Ogi Networks had inadequate password controls and account audits, while Data Cymru lacked a compliant data-processing agreement with its provider and failed to establish clear responsibilities for monitoring and reviewing access. Both organisations were found to have breached UK GDPR security requirements.
- Following the incident, both organisations strengthened access controls and introduced more regular account reviews, while Data Cymru and Ogi Networks signed a compliant data-processing agreement in April 2025.
United States
OpenAI introduces locked chats for ChatGPT sensitive conversations
- OpenAI introduced a privacy feature that allows users to conceal sensitive ChatGPT conversations from their regular chat history, pinned chats and search results. The feature is designed to protect private discussions when users share their devices or use ChatGPT in public places.
- This feature moves selected conversations into a dedicated section in the sidebar. Users must set a four-digit PIN to access them, with mobile devices also supporting facial recognition or fingerprint authentication.
- Locked conversations have a separate memory space, preventing new information from those chats from updating the user’s regular ChatGPT memory. However, the feature does not provide end-to-end encryption, and OpenAI may still access conversations or use them for model training if the user has enabled the relevant data-sharing setting.
- The feature is available across Free, Go, Plus, Pro, Business, Enterprise and Edu plans. However, users who forget their PIN and reset it will permanently lose their locked chats and associated files.
Meta faces $25 billion fine over data privacy practices
- A New Mexico judge ordered Meta to pay $25 billion after a jury found that the company repeatedly misled the public about how Facebook user data was handled and shared. The judge said the misconduct was deliberate and continued for more than a decade, from 2010 to 2021.
- This arose partly from the Cambridge Analytica scandal, in which Facebook user data was used for political consulting, including work connected to the 2016 US presidential election. Jurors also found that Meta misrepresented how it profited from content involving hate speech and how it enforced its policies.
- The court found that Meta had violated state laws more than 40 million times, with communications involving senior executives, including Mark Zuckerberg, described in the ruling as intended to mislead the public. Meta rejected the accusations, argued the case threatened online speech, and said it would appeal the decision.
Europe
European Commission investigates Booking.com over fraudulent listings and scams
- The Commission has launched an investigation into Booking.com over concerns that the platform has failed to adequately protect users from fake accommodation listings and phishing scams. Regulators suspect its safeguards may not satisfy the EU Digital Services Act, which allows penalties of up to 6% of a company’s global annual turnover.
- Investigators will examine whether Booking.com properly assesses fraud risks, prevents misuse of customer data and clearly discloses when accommodation providers pay to increase their visibility. The inquiry follows reports of scammers creating fake properties and hijacking genuine reservations to trick travellers into sending money.
- Reports of fraud linked to the platform have increased, with the Dutch consumer association recording 200 complaints involving €65,000 in losses during 2025, compared with 89 reports totalling €25,000 the previous year. Booking.com has also previously reported a rise in scams, partly attributed to criminals using generative AI to produce convincing fraudulent messages and booking links.
- Booking.com said it is cooperating with regulators and considers security and customer protection priorities.
Greece bans social media for children under 15
- Greece’s parliament has passed legislation prohibiting children under 15 from using social media, with the restrictions scheduled to take effect on 1 January 2027.
- Greek Prime Minister said platforms have roughly two and a half months to develop systems that can confirm users’ ages while safeguarding their privacy.
- Greece is working with six other EU countries to find an age-verification approach in coordination with the European Commission.
International
Critical PHP vulnerabilities lead to global government website outages
- Government agencies and organisations across Europe took websites and online services offline after researchers uncovered critical vulnerabilities in PHP, a programming language used by a large proportion of the web.
- The vulnerabilities potentially affect millions of internet-facing systems, including websites built on platforms such as WordPress, Drupal, Joomla and Magento. Data from internet intelligence firm Censys indicated that around 1.5 million of 2.6 million identified hosts were running PHP versions without an available fix.
- Cybersecurity researchers independently identified the flaws using AI-assisted tools, including systems from Google and OpenAI. Amazon also reported detecting attempted exploitation before public disclosure, followed by further activity, raising concerns that AI could help attackers identify and exploit vulnerable servers more quickly.
- The incident caused widespread precautionary shutdowns, including services operated by cybersecurity agencies, government departments, hospitals and other organisations in the Netherlands, France, Denmark and Finland.
For the latest updates on PHP vulnerabilities, locked chats for ChatGPT, social media ban for under 16s, and other major data privacy news, visit our Data Protection News Hub.



