Study: Internet Freedom Declines In US, Germany

A woman walks past an illustration of internet at the expo of the World Internet Conference in Wuzhen town of Tongxiang city, Zhejiang province, China November 8, 2025. REUTERS/Tingshu Wang
A woman walks past an illustration of internet at the expo of the World Internet Conference in Wuzhen town of Tongxiang city, Zhejiang province, China November 8, 2025. REUTERS/Tingshu Wang
TT

Study: Internet Freedom Declines In US, Germany

A woman walks past an illustration of internet at the expo of the World Internet Conference in Wuzhen town of Tongxiang city, Zhejiang province, China November 8, 2025. REUTERS/Tingshu Wang
A woman walks past an illustration of internet at the expo of the World Internet Conference in Wuzhen town of Tongxiang city, Zhejiang province, China November 8, 2025. REUTERS/Tingshu Wang

Internet freedom declined in both the United States and Germany as Western democracies joined authoritarian states in imposing growing restrictions online, an annual survey by Freedom House said Thursday.

The Washington-based democracy promotion research group said global internet freedom fell for the 15th straight year, with declines in a number of countries that remain classified as "free."

"We find worsening repression in authoritarian and authoritarian-leaning states, largely because governments in those countries see restrictions on the internet and online expression as a means of maintaining power," said Kian Vesteinsson, the report's co-author.

"Much more specific to 2025, we've seen conditions in democracies declining," he told AFP.

"Unfortunately we see across North America and Western Europe generally a trend towards closing civic space in some countries, and in others deepening restrictions on people posting hateful or problematic content," he said.

The United States scored 73 on a scale of 100 on internet freedom in the year through May 2025 covered by the report, its lowest-ever figure and down three points from the previous year.

The report pointed in part to the detention by President Donald Trump's administration of a number of non-US citizens over their online expression.

Secretary of State Marco Rubio has vowed to deport people over statements on Israel, decisions challenged in courts.

Germany also saw a decline of three points to 74. Freedom House said that Germany was seeing growing self-censorship and also the zealous enforcement of laws that prohibit hate speech and defamation.

It pointed to a suspended jail sentence and fine imposed on the editor of a far-right website over a social media post that included a manipulated image to criticize a politician.

The Trump administration has repeatedly criticized the free speech record of Germany, which says its Nazi past makes it critical to impose safeguards.

The most severe declines in the report were registered by Kenya, which briefly shut down the internet in response to nationwide protests, as well as Venezuela and Georgia.

Two countries were downgraded -- Serbia was classified as "partly free" instead of "free" and Nicaragua was marked as "not free" instead of "partially free."

Bangladesh saw the biggest gain, as a new government formed after a student revolt loosened restrictions in the country.

Freedom House, founded during World War II to promote democracy, was historically funded largely by the US government but operated independently.

Trump on returning to office slashed funding to rights groups including Freedom House, which has laid off staff.



Microsoft Cuts OpenAI Revenue Share in a Fresh Step to Loosen Their AI Alliance

FILE PHOTO: A Microsoft logo is seen next to a cloud in Los Angeles, California, US June 14, 2016. REUTERS/Lucy Nicholson/File Photo
FILE PHOTO: A Microsoft logo is seen next to a cloud in Los Angeles, California, US June 14, 2016. REUTERS/Lucy Nicholson/File Photo
TT

Microsoft Cuts OpenAI Revenue Share in a Fresh Step to Loosen Their AI Alliance

FILE PHOTO: A Microsoft logo is seen next to a cloud in Los Angeles, California, US June 14, 2016. REUTERS/Lucy Nicholson/File Photo
FILE PHOTO: A Microsoft logo is seen next to a cloud in Los Angeles, California, US June 14, 2016. REUTERS/Lucy Nicholson/File Photo

Microsoft said Monday it will no longer pay a share of its revenue to ChatGPT maker OpenAI, the latest move to untether a close partnership that helped unleash an artificial intelligence boom.

OpenAI relied exclusively on Microsoft's investments in cloud computing services to build the technology that helped make ChatGPT a household name. Microsoft, in turn, relied on OpenAI's technology to build its own AI assistant Copilot.

But the partnership has evolved as San Francisco-based OpenAI, founded as a nonprofit, has shifted to a capitalistic enterprise on a path toward an initial public offering on Wall Street and has balanced its reliance on Microsoft with other cloud partners like Amazon, Google and Oracle, The AP news reported.

OpenAI said Monday it will continue to pay Microsoft a share of its revenue through 2030.

The two companies said Microsoft remains the primary cloud computing partner for OpenAI, and products made by the AI company will ship first on Microsoft's cloud platform, called Azure, “unless Microsoft cannot and chooses not to support the necessary capabilities.”

Wedbush Securities analyst Dan Ives said in a note to investors Monday that the new agreement “puts OpenAI on a strong path forward to going public through IPO given its clearer opportunity in the cloud environment while reducing significant barriers from its original partnership with Microsoft.”

Ives said it's also important for Microsoft as it “looks to develop tech independence from OpenAI” in advancing Copilot's capabilities and partnering with other AI providers such as OpenAI rival Anthropic, maker of the chatbot Claude.


China Blocks Meta from Acquiring AI Startup Manus

The smart assistant 'Manus' on a smartphone screen (AFP)
The smart assistant 'Manus' on a smartphone screen (AFP)
TT

China Blocks Meta from Acquiring AI Startup Manus

The smart assistant 'Manus' on a smartphone screen (AFP)
The smart assistant 'Manus' on a smartphone screen (AFP)

China on Monday blocked US tech giant Meta’s acquisition of the artificial intelligence startup Manus, in an unexpected move to reverse a deal that apparently aroused Beijing's concerns about the transfer of advanced technology.

In a one-line statement, China’s National Development and Reform Commission, the country's top planning agency, said it was prohibiting the foreign acquisition of Manus and had required all the parties to withdraw from the deal. It did not specifically name Meta Platforms, which owns Facebook and Instagram, The AP news reported.

Manus, which has Chinese roots but is based in Singapore, provides a general-purpose AI agent that can autonomously carry out sophisticated tasks like coding an app, doing market research or preparing quarterly budgets.

The decision was made by the commission’s Office of the Working Mechanism for Security Review of Foreign Investment in accordance with Chinese laws and regulations, the statement said. It came after Chinese authorities said they were looking into the deal earlier this year.

The commission did not elaborate on the reasons for the ban. The announcement came less than a month before US President Donald Trump's planned visit to Beijing to meet Chinese leader Xi Jinping in May.

Meta announced in December that it was acquiring Manus, in a rare case of a major US tech group buying an AI company with strong links to China. Its deal with Manus was expected to help expand AI offerings across Meta’s platforms.

Meta had said there would be “no continuing Chinese ownership interests in Manus” and that Manus would discontinue its services and operations in China. But China said in January that it would investigate whether the acquisition would be consistent with its laws and regulations.

China’s commerce ministry said at the time that any enterprises engaging in outward investment, technology exports, data transfers and cross-border acquisitions must comply with Chinese law. Meta had said most of Manus’ employees were based in Singapore.

Before the deal, Manus’ parent was Singapore-based Butterfly Effect Pte, but the AI startup traces its roots back to Beijing-registered entities with similar names that were established several years earlier.

Manus did not respond to a request for comment. Its website says the company “is now part of Meta," indicating that the deal had already been completed.

Meta said on Monday that the Manus transaction “complied fully with applicable law.”

“We anticipate an appropriate resolution to the inquiry,” the California-based company said in a statement.

Analysts said the decision is a sign that China’s communist leaders are tightening scrutiny of the AI industry amid intensifying geopolitical rivalry with the US over the technology.

“China is showing the world that it is willing to play hardball when it comes to AI talents and capabilities, which the country views as a core national security asset,” said Lian Jye Su, chief analyst at the technology research and advisory group Omdia. “It is strongly indicative of what Chinese authorities may do going forward regarding acquisitions involving Chinese deep-tech companies.”

Beijing’s acquisition ban could deter similar acquisition plans by US tech giants going forward, he said. “In the context of rivalry, it mirrors US export controls, entity lists, and investment curbs on China,” said Su.

Meta’s interest in Manus reflects a broader tech industry race to lead in the development of AI agents that can go beyond a chatbot’s capabilities to take computer-based actions on people’s behalf.

Meta last month acquired Moltbook after it attracted viral attention as a social network built for AI agents to make posts and interact with each other. That was after OpenAI, maker of ChatGPT, hired the creator of AI agent OpenClaw, formerly called Moltbot and the technology upon which Moltbook was built.


Google to Build AI Campus in South Korea, Presidential Office Says

South Korean President Lee Jae Myung (R) shakes hands with Demis Hassabis (L), co-founder and CEO of Google DeepMind and the architect behind the AlphaGo artificial intelligence system, during their meeting at the presidential office in Seoul, South Korea, 27 April 2026. (EPA/Yonhap)
South Korean President Lee Jae Myung (R) shakes hands with Demis Hassabis (L), co-founder and CEO of Google DeepMind and the architect behind the AlphaGo artificial intelligence system, during their meeting at the presidential office in Seoul, South Korea, 27 April 2026. (EPA/Yonhap)
TT

Google to Build AI Campus in South Korea, Presidential Office Says

South Korean President Lee Jae Myung (R) shakes hands with Demis Hassabis (L), co-founder and CEO of Google DeepMind and the architect behind the AlphaGo artificial intelligence system, during their meeting at the presidential office in Seoul, South Korea, 27 April 2026. (EPA/Yonhap)
South Korean President Lee Jae Myung (R) shakes hands with Demis Hassabis (L), co-founder and CEO of Google DeepMind and the architect behind the AlphaGo artificial intelligence system, during their meeting at the presidential office in Seoul, South Korea, 27 April 2026. (EPA/Yonhap)

South Korea and ‌Google have agreed to build an artificial-intelligence campus in Seoul to develop cooperation between the tech firm and local engineers and startups, Kim Yong-beom, a presidential policy adviser, said on Monday.

South Korean President Lee Jae Myung met with Google DeepMind Chief Executive Officer Demis Hassabis in Seoul on Monday, with the Science Ministry and the company signing a memorandum of understanding on the campus, Kim said.

South Korea requested Google send ‌at ⁠least 10 engineers to the ⁠AI campus from Google's headquarters in the United States and Hassabis said he would consider that, Kim said.

The Google AI campus will be the first of its kind in the world for the US company, the presidential adviser said.

President Lee ⁠and Hassabis shared their thoughts about ‌the outlook for AI and ‌its impact on people, Kim said.

Lee raised the need ‌for the introduction of a base wage ‌in case of job losses caused by AI at the meeting.

Hassabis said he hoped with this partnership "to help with training up the next generation in these amazing technologies through ‌internships at our AI hub and other training programs."

DeepMind would like to deepen ⁠partnerships with ⁠Korean companies from Samsung and SK Hynix to Hyundai's Boston Dynamics and LG and "instigate new joint projects" with them, Hassabis said.

He described South Korea as a "great industrial base" in all of the key AI areas, from chips to robotics.

The historic match between DeepMind's AlphaGo program and Go player Lee Sedol in Korea a decade ago signaled the beginning of the modern AI era and inspired many advances in AI, including its work in science like the Alphafold system for protein folding, Hassabis said.