Italy, 2 others ban DeepSeek; OpenAI responds with o3-mini

Since the launch of its artificial intelligence (AI) chatbot in January, DeepSeek has dominated the tech sector, with Western firms scrambling to grasp how an unknown Chinese startup had become a global phenomenon overnight. Industry leader OpenAI was quick to respond by launching o3-mini, its most cost-efficient reasoning model.

DeepSeek is also proving to be a headache for regulators. While the Trump administration weighs a restriction to protect American companies, the Italian government is moving fast, banning the Chinese company for alleged opaque use of Italians’ data. Taiwan has implemented a partial ban, and nearly a dozen other nations in Europe and Asia are mulling similar measures.

OpenAI’s response to DeepSeek: o3-mini

OpenAI announced the launch of o3-mini on Friday, describing it as “the newest, most cost-efficient model in our reasoning series.”

First previewed last December, the o3-mini is the latest member of the AI giant’s ‘o’ reasoning models—the first was o1, which launched early 2024, but the company reportedly skipped o2 due to potential trademark infringements. Unlike GPT-4o, which focuses on mass tasks and is more creative, the ‘o’ family of models is more geared toward complex and structured tasks.

OpenAI says the new model is optimized for science, mathematics, and coding, all while reducing the latency that prior models faced.

More importantly, it offers these advantages while maintaining low costs. This is a direct response to DeepSeek, whose claim to fame was its cost-effectiveness. While OpenAI reportedly spent hundreds of millions of dollars to train its models, DeepSeek claimed to have spent less than $6 million to achieve the same results.

OpenAI has priced o3-mini at $0.55 and $4.40 per 750,000 input and output words, respectively, which is around a third of the cost of the previous model. However, it’s still higher than DeepSeek, which charges $0.14 and $2.19 for similar input and output words, respectively.

“The release of o3-mini marks another step in OpenAI’s mission to push the boundaries of cost-effective intelligence […] As AI adoption expands, we remain committed to leading at the frontier, building models that balance intelligence, efficiency, and safety at scale,” the company stated.

o3-mini is available to all ChatGPT users, marking the first time that free users get to try out the company’s reasoning models, in yet another direct response to DeepSeek’s mass market appeal. It will be embedded in the ChatGPT chatbot under the “Reason” feature. However, paying users will unlock extra features, which OpenAI says include more intelligent responses and higher message limits. To get unlimited access to the new model, users will need to pay $200 monthly for ChatGPT Pro.

DeepSeek spooks regulators—bans in Italy, Taiwan, Texas

Since it launched its chatbot, which became wildly popular globally, DeepSeek has unsettled Western regulators, prompting them to respond with restrictions and bans.

On Friday, Italy’s data protection authority, Garante, banned the Chinese firm’s chatbot, pointing to a lack of transparency on how it would use data collected from Italian users. Garante claimed to have sent DeepSeek a series of questions seeking more information about how it collects, stores, and uses the data, and it wasn’t satisfied with the responses.

It’s not the first time that Garante has cracked down on an AI model. In April 2023, the watchdog banned ChatGPT over data privacy concerns and launched an investigation into whether OpenAI had breached the European General Data Protection Regulation (GDPR). However, less than a month later, it lifted the ban and stated that OpenAI had addressed the concerns.

While Italy is among the first to fully ban DeepSeek, others, like Taiwan, are restricting its use in more specific areas. On Monday, Taiwanese Prime Minister Cho Jung-tai banned the use of the AI model in the public sector to “ensure the country’s information security” is adequately protected.

Additionally, Taiwan is concerned about its citizens’ data ending up in Chinese hands owing to escalating tensions between the two over China’s ramped-up pressure for unification. PM Jung-tai also expressed concerns that the Chinese government might use the AI model to enforce censorship, with Beijing believed to have unfettered access to all Chinese AI models.

And then there’s the United States, upon which the Western world awaits direction on how to respond to DeepSeek’s overnight dominance. Many American leaders in the political, tech and financial sectors have called on the Trump administration to move fast and ban the Chinese model. OpenAI, which stands to lose the most, has even accused DeepSeek of improperly using its models to train its AI, a claim Trump’s AI czar David Sacks backed.

As Trump considers his next move, Texas isn’t sitting idly by and has banned the use of DeepSeek on any government devices.

“Texas will not allow the Chinese Communist Party to infiltrate our state’s critical infrastructure through data-harvesting AI and social media apps,” Governor Greg Abbott stated.

In order for artificial intelligence (AI) to work right within the law and thrive in the face of growing challenges, it needs to integrate an enterprise blockchain system that ensures data input quality and ownership—allowing it to keep data safe while also guaranteeing the immutability of data. Check out CoinGeek’s coverage on this emerging tech to learn more why Enterprise blockchain will be the backbone of AI.

Watch: Demonstrating the potential of blockchain’s fusion with AI