Yazılar

Italy’s Data Regulator Blocks DeepSeek AI Chatbot Over Privacy Concerns

/in /tarafından

Italy’s data protection authority, the Garante, has ordered Chinese AI startup DeepSeek to block its chatbot in the country after the company failed to address concerns over its privacy practices. The regulator had questioned DeepSeek about its handling of personal data, including details on what data is collected, its sources, the purposes for which it is used, its legal basis, and whether it is stored in China.

The Garante’s decision came after the company provided what was deemed “totally insufficient” information, prompting the watchdog to take immediate action to protect Italian users’ privacy. DeepSeek has yet to comment on the ruling.

The Chinese startup, which recently claimed that its AI models rival or outperform industry-leading U.S. models at a fraction of the cost, has made headlines for surpassing ChatGPT as the top-rated free app on Apple’s App Store in the U.S. However, its swift rise has drawn increasing scrutiny over data protection.

The Garante’s order, which took effect immediately, also includes an investigation into DeepSeek’s data handling practices. Meanwhile, other European regulators, including those in France and Ireland, are investigating the chatbot’s privacy policy.

In response to Italy’s concerns, DeepSeek stated that it had removed its AI assistant from Italian app stores after facing scrutiny. However, Agostino Ghiglia, a member of the Garante board, revealed that the company’s stance—which claimed it was not subject to Italian regulation—further aggravated the situation, prompting the block. Ghiglia emphasized that DeepSeek’s lack of cooperation had made the situation worse.

As of Friday, some Italian users who had previously downloaded the app reported that the chatbot was still functional on their devices, and the web version of the service remained operational. The Garante emphasized that citizens must have the right to consent based on how their data is handled, especially when servers in countries outside the EU, such as China, may not provide the same privacy guarantees as European standards.

The Garante has been at the forefront of data protection in Europe, with a history of taking action against companies, including a brief ban on Microsoft-backed ChatGPT two years ago over potential privacy rule violations.

 

Microsoft and Meta Defend Heavy AI Investments Despite DeepSeek’s Low-Cost Advantage

/in /tarafından

In response to the breakthrough low-cost AI models developed by Chinese startup DeepSeek, CEOs of Microsoft and Meta have defended their substantial investments in artificial intelligence, emphasizing that the heavy spending is essential to staying competitive in the rapidly growing field. DeepSeek’s claims of outperforming Western AI models at a fraction of the cost have sparked concerns over the U.S. tech industry’s dominance, but both executives stressed that building extensive computing infrastructures is crucial to meeting rising corporate demands.

Meta CEO Mark Zuckerberg highlighted the strategic advantage that heavy investments in capital expenditure and infrastructure will bring over time. Microsoft CEO Satya Nadella echoed this sentiment, stating that such investments are needed to address the capacity constraints that have limited the company’s ability to capitalize fully on AI opportunities. Nadella also noted that as AI becomes more efficient and accessible, demand for the technology will grow exponentially.

Microsoft has allocated $80 billion for AI in its current fiscal year, while Meta has committed up to $65 billion. This stands in stark contrast to the roughly $6 million that DeepSeek claims to have spent on developing its AI model. However, U.S. executives and analysts note that DeepSeek’s reported costs are limited to computing power, not including broader development expenses.

Despite these substantial investments, investor patience is waning. Microsoft shares dropped 6% after the company revealed that its Azure cloud business growth would fall short of third-quarter expectations. Brian Mulberry, portfolio manager at Zacks Investment Management, emphasized the need for a clearer path to monetizing the investments.

Meanwhile, Meta’s stock rose more than 4% following a strong fourth-quarter performance, though its first-quarter sales forecast was underwhelming. Analysts, like Daniel Newman from Futurum Group, pointed out the disparity between capital expenditure and revenue generation in the AI sector.

Both companies have indicated efforts to moderate spending. Microsoft CFO Amy Hood stated that capital expenditures for the third and fourth quarters would remain around $22.6 billion, similar to the previous quarter, with growth rate expectations for fiscal 2026 being lower than in fiscal 2025.

 

French Privacy Watchdog to Investigate DeepSeek Over AI and Data Protection

/in /tarafından

France’s data privacy authority, the CNIL, announced on Thursday that it will question DeepSeek to assess the workings of its AI system and potential privacy risks for users. The Chinese AI startup gained international attention after revealing that training its DeepSeek-V3 model required less than $6 million in Nvidia H800 computing power.

A CNIL spokesperson confirmed that its AI department is currently analyzing DeepSeek’s tool and will engage with the company to understand its system and data protection measures. The French regulator is among the most active in Europe, having previously fined tech giants like Google and Meta for privacy violations.

DeepSeek is also under scrutiny in other parts of Europe. Italy’s data protection authority recently requested details on its handling of personal data, while Ireland’s Data Protection Commission has inquired about data processing practices related to Irish users.

The European Union maintains strict privacy protections under its General Data Protection Regulation (GDPR), widely regarded as one of the world’s most comprehensive data privacy laws. GDPR violations can result in fines of up to 4% of a company’s global revenue. Additionally, new EU AI regulations impose transparency obligations on high-risk AI models, with penalties ranging from 7.5 million euros (or 1.5% of turnover) to 35 million euros (or 7% of global turnover), depending on the severity of violations.

As regulatory scrutiny intensifies, DeepSeek faces mounting pressure to demonstrate compliance with European data protection standards.