A Microsoft outage lasting almost 24 hours affected millions of users in 74 countries, coinciding with the EU’s first summit on artificial intelligence in Dublin, according to Politico Europe. The disruption impacted many officials attending the summit, prompting renewed debate about Europe’s dependence on non-European digital infrastructure.
Arthur Mensch, chief executive of French AI firm Mistral, told leaders that only 20% of European digital services use European infrastructure, arguing that public spending should support local providers. French President Emmanuel Macron and several European politicians called for greater digital independence, with some urging government contracts to favour European technology. Others, such as the Digital SME Alliance, said the outage showed the risk of relying on a small number of global tech giants.
The incident follows similar concerns raised after the CrowdStrike outage in 2024. This time, the European Commission has a draft cloud law under discussion that would rank cloud providers by their level of sovereignty. The outage may increase pressure on Brussels to act.
Uber has told drivers that hackers stole their personal and driver licence data from Fireworks AI, a vendor that provides access to open AI models, according to TechCrunch. The breach, confirmed by Uber, began when hackers stole internal credentials from Fireworks, which were then used to access Uber’s data. Notices filed with US state attorneys general list at least 7,870 affected in Massachusetts, South Carolina and Rhode Island, with more expected as California and Illinois residents have also been notified.
Fireworks, valued at $17.5bn in July, reported two incidents in October: one involving stolen credentials and another where attackers accessed its development machines through a cloud provider breach. Fireworks says billing data, secrets, datasets and unfinished model weights were taken. Notion, another customer, said fewer than 1% of its AI chat transcripts were exposed. Fireworks attributed the attack to a group that sells stolen AI access, a practice called LLMjacking.
The presence of driver licence data with an AI inference provider raises questions about data flows and vendor oversight. Uber referred further queries to Fireworks, which has not answered detailed questions. This is Uber’s second breach this year, following an incident at Uber Freight in August.
Firms relying on AI vendors should review how sensitive data is shared and stored, and consider the risks of credential theft and supply chain attacks.
Italy’s data protection authority has fined Meta €805m after finding that the Facebook app for Android accessed photos, videos and their metadata on Italian users’ phones without proper consent or clear notice. The Garante said Facebook’s gallery suggestions feature, available since 2023, prompted users to grant access to their phone’s gallery through a single, vague permission, making it unclear that gallery suggestions would be enabled.
Meta argued the case should be handled by Ireland’s regulator under GDPR, but the Garante used ePrivacy rules, which do not allow a one-stop shop. The regulator said about 20 million Android users in Italy were affected. It also cited Meta’s poor cooperation and a 2019 decision against the company. Meta’s changes after the complaint were seen as too late and only applied to users who updated the app.
Meta must now ask for specific consent for gallery suggestions and keep the feature off by default, even for users who previously granted access. The probe did not cover whether gallery content was used for AI training, which Meta denies. The Garante will separately examine a new cloud processing feature launched in 2026.
The fine is among the largest in Europe for data privacy breaches. The case signals stricter enforcement of consent requirements for digital platforms operating in Italy and the EU.
HubSpot has reinstated co-founder Brian Halligan as chief executive, replacing Yamini Rangan with immediate effect. The company described the move as a “refounding” as Halligan aims to overhaul product, culture and sales. Halligan, who led HubSpot from its founding in 2005 through its 2021 stock market peak, returns after a period at Sequoia Capital. Rangan, who oversaw customer growth and rising revenue, will remain until December for the handover.
The leadership change follows a year in which HubSpot’s shares lost half their value, driven by slowing growth and concerns over competition from AI-native tools. Halligan told Bloomberg the company had focused too much on short-term financial targets at the expense of product quality. He said he wants to rebuild and reimagine the business rather than see it acquired or diminished.
HubSpot is also making changes in its finance team. Kate Bueker will retire as CFO in 2027, with Karen Burgess from Adobe set to take over. The company recently announced 660 job cuts as it pivots to AI, and reaffirmed its financial guidance for the year.
Many software firms are facing similar leadership shifts, as the sector adjusts to AI and market volatility. HubSpot’s next steps on product and sales strategy will be closely watched by industry leaders and investors.
Microsoft will turn on Anthropic’s Claude AI models by default for UK Copilot customers from 9 December, The Register reported. Anthropic, acting as a Microsoft subprocessor, will process customer data in the US. UK organisations that do not select a setting by the deadline will have Claude enabled, though administrators can still opt out. Claude is already the default for most Microsoft commercial customers, but the EU, EFTA and UK had been exceptions until now.
Microsoft’s Copilot documentation states that Claude models fall outside the EU Data Boundary and are not covered by in-country processing commitments. This means UK customer data will be handled under US jurisdiction, which may raise compliance concerns for banks, insurers and healthcare providers. Some advanced Anthropic models, such as Claude Fable 5 and Claude Mythos 5, remain off by default everywhere, and require explicit opt-in.
The move comes as the Institute for Public Policy Research (IPPR) warns about the UK public sector’s reliance on US technology. Only 15.4% of spending with top tech suppliers stays with British-owned companies, while the US receives the largest share. IPPR urges the government to act as a single buyer and improve supplier flexibility.
UK administrators must confirm their Copilot settings by 9 December to avoid defaulting to Claude. The shift may prompt further scrutiny of cross-border data flows and supplier concentration in regulated industries.
Ten AI companies, including OpenAI, Anthropic, Microsoft and Mistral, have signed the Paris Call for AI Safety and Security at a Paris summit hosted by President Emmanuel Macron. The declaration, also signed by six countries, sets out commitments to improve independent risk evaluation, share incident alerts, and allow expert testing of advanced AI models. Signatories have agreed to develop fail-safe mechanisms to shut down models and publish non-sensitive test results, but major players like Google, Meta and China have not joined.
The agreement follows earlier safety initiatives and is open to further signatories, but sets no deadlines or binding risk thresholds. French AI safety group CeSIA, which signed the declaration, criticised the lack of clear obligations, enforceable standards and government authority to suspend risky models. CeSIA called for public authorities to set the rules and for a timetable to be established.
For regulated firms, the declaration signals growing pressure for independent oversight and transparency in AI deployment. The absence of binding rules and the limited group of signatories mean regulatory expectations may shift further as governments and industry debate enforcement and access. The framework for shared testing and incident reporting is still to be developed.
Macron also called for pre-certification of AI tools used by children, likening it to toy safety checks, and warned of the mental health risks posed by unregulated AI systems for young users. A charter on safer AI for children was expected to follow.
Tencent’s chief AI scientist, Yao Shunyu, warned that AI models will soon be able to cause harm in biology, as they already can in cybersecurity, according to the South China Morning Post. Speaking at a science event in Hong Kong, Yao said Chinese firms releasing open models must take safety seriously, as biology risks may be harder to manage than cyber risks.
Tencent recently launched its new flagship AI model, Hy4, which the company said was trained on defensive security data and has strong basic biology knowledge. However, Tencent noted Hy4 is weaker than leading models at designing experiments. The company plans to release Hy4’s weights and a safety testing report soon, according to Reuters. Benchmarking firm Artificial Analysis ranks Hy4 as the second-best Chinese model, the SCMP said.
Yao pointed to US firms like Anthropic as examples of keeping top models closed, but suggested a mix of open and closed models could be effective, as open models can also help defenders. This debate is likely to influence future regulatory and safety standards for AI in sensitive sectors.
The US Justice Department has created a Super Intelligence Task Force to lead its work on AI crimes and lawsuits, the department said. The group will target foreign states, criminal gangs, and individuals or firms using AI for fraud. Associate Attorney General Stanley Woodward and Associate Deputy Attorney General Stephen Weiss will head the task force.
The new group will absorb the DOJ’s AI Litigation Task Force, which was set up in January to challenge state AI laws in federal court, according to an internal memo cited by Nextgov. The task force will also set priorities and write rules for prosecuting AI crimes, and act as the DOJ’s main contact with the White House’s Super Intelligence Force.
Attorney General Todd Blanche stated the department will keep pursuing those who use AI unlawfully. Weiss added that the DOJ would urge firms to maintain strong controls to ensure their AI systems operate as intended. The department did not disclose the staffing or budget for the new task force.
Firms in regulated sectors should watch for new compliance requirements and enforcement actions as the DOJ sharpens its focus on AI-related risks.
US prosecutors have charged Jacob Soall, Sharon Hazan and Guy Sheetrit with running a $129m fraud that used AI to mimic CEOs’ voices and images. The men allegedly called employees, posed as company chiefs, and directed them to transfer large sums for supposed secret deals. Victims included a Missouri hospital system, which lost $4.8m, and a UK fintech, which lost $11.6m. Funds were sent to accounts in several countries and converted to cryptocurrency, according to the indictment.
The group used AI voice cloning to make the impersonations convincing and involved a fake lawyer to provide bank details. Soall mainly posed as UK lawyers and kept a script for the calls, while Hazan and Sheetrit are accused of laundering the proceeds. Hazan has a prior conviction for bank fraud conspiracy. The men face charges including wire fraud and aggravated identity theft, with maximum sentences of up to 20 years. The charges are allegations and the men are presumed innocent.
The case follows a 2019 incident where criminals used voice cloning to steal $243,000. Since then, AI-enabled impersonation scams have increased. The incident raises new concerns for regulated firms about the adequacy of controls and staff training against AI-driven fraud.
Microsoft is ending its 2023 deal to rent AI computing power from Oracle, according to Bloomberg. The company has been switching off the rented capacity in recent months after deciding not to renew the contract, with the servers running older Nvidia chips. Microsoft is reviewing more than $80B in contracts with outside cloud providers as it updates its AI infrastructure.
The Oracle deal supported Microsoft's AI-powered Bing search, but the hardware no longer matched its needs. Gartner has warned that ageing AI infrastructure could lead to significant write-downs as newer chips become available. Attempts to expand Microsoft’s use of Oracle capacity collapsed in June, and further negotiations with other suppliers such as Nscale have also stalled.
Oracle faces increasing pressure as demand shifts. Its heavy spending on AI data centres has pushed free cash flow negative and left its credit rating just above junk. Microsoft’s decisions signal a focus on cost control and up-to-date technology, affecting suppliers’ revenue and market positions.
Senior executives in regulated sectors should monitor supplier stability and contract terms as cloud infrastructure needs evolve rapidly. The next moves by Microsoft and other large buyers could reshape the AI infrastructure market.