Chat GPT Ban in Hong Kong and China (Updated)

Chat GPT Ban (Openai’s services are not available in your country)

Openai's services are not available in your country is the error that you are likely to see if you are in Hong Kong and try to access Chat GPT. The recent Chat GPT ban blocking access to its services not only for casual users but also for developers in mainland China and Hong Kong has surprised the AI developer community in Hong Kong and China. This move, which took effect on July 9, 2024, halted all API access for these regions, forcing developers to seek alternatives. Despite the immediate challenges, some AI experts believe this could help to accelerate the growth of China and Hong Kong's domestic AI industry, this is a likely and unintended consequence for Open AI.

Impact on Developers and Industry Response

The OpenAI Chat GPT ban impacts numerous Chinese companies and developers who rely on Chat GPT generative AI models. Many have been using VPNs and third-party applications to bypass restrictions and gain access to OpenAI's Chat GPT API services. With the ban, these developers are now compelled to move towards to local AI solutions. In response, Chinese AI firms have swiftly adapted. Beijing-based Zhipu AI, for instance, launched a “special house-moving plan” offering free tokens and migration support to lure developers from Chat GPT to their own platforms. Major players like Alibaba, Baidu, and AI start-ups like Baichuan and 01.ai have also introduced incentives, such as steep discounts and technical support, to attract users to their AI models.

Growing Competition and Local AI Development

China already boasts over 200 home-grown large language models similar in capability to Chatgpt (LLMs), with 117 approved for public release as of March 2024. This local innovation is expected to intensify as companies scramble to fill the void left by OpenAI. The competitive landscape has even sparked a price war among Chinese AI firms, with some offering free access to less powerful models to gain market share. For example, SenseTime is a leading AI software company with its HQ in Hong Kong. It was founded 2014 by Tang Xiao'ou, a professor at the Chinese University of Hong Kong. This local AI company has grown to become a leading AI firm, known particularly in the fields of computer vision, including facial and image recognition and object detection.

Geopolitical Context and Future Prospects

This development is part of a broader tech war between the US and China, with the Biden administration imposing restrictions on China's access to advanced semiconductors and AI technologies. OpenAI’s action is seen as another step in this ongoing conflict, potentially driving further advancements in China’s self-sufficiency in AI technology. While this Chat GPT ban poses significant short-term challenges for Chinese developers, it may ultimately catalyse the growth and sophistication of China’s AI sector, pushing it closer to self-reliance and innovation.

Confusion Amid Contradictory Measures

Interestingly, Chat GPTs measures seem contradictory, given that Apple plans to launch its Chat GPT powered updates to iOS 18 later this year. Many of these features were just showcased in the Apple event marking the launch of the new iPhone 16. Furthermore, applications like Poe and others available in the Hong Kong App Store already provide access to not only Chat GPT 40 but other major ones such as Claude 3.5 and Llama from Meta. Additionally, Microsoft still allows access to OpenAI CoPilot, further complicating the landscape. These discrepancies highlight the complex and often conflicting nature of tech policies in the current geopolitical climate, raising questions about the efficacy and consistency of such restrictions which seem increasingly futile and pointless.

Driving Toward Global Competitors

The OpenAI ban could inadvertently drive Hong Kong and China-based companies towards the increasing number of global competitors available. With more than 200 home-grown LLMs and growing support from other tech giants like Microsoft, which continues to allow access to OpenAI CoPilot, developers have numerous options to choose from. This move might diversify and strengthen the global AI ecosystem, fostering competitive innovation. This decision is particularly frustrating for developers and companies in China and Hong Kong. Many have invested significant time and resources into integrating OpenAI’s models into their services. The abrupt cutoff not only disrupts ongoing projects but also undermines the trust and stability expected from such partnerships.

PTS Opinion

By driving developers to alternative platforms, OpenAI and Chat GPT risk losing a significant user base and revenue from one of the largest tech markets in the world, how many times will users see openai's services are not available in your country and eventually just forget about it and opt for a local alternative. The inconsistency in policy enforcement—where some platforms and services remain accessible or in some cases not directly but through 3rd party services or VPN's, only adds to the confusion and frustration, making the ban seem arbitrary and ineffective. Rather than stifling innovation, these measures may inadvertently spur a surge in local AI development, as Chinese companies strive to fill the gap left by Chat GPT.

Update (September 2024)

Availability of Azure OpenAI Services in Hong Kong

Azure OpenAI Services continue to be available in Hong Kong, providing businesses, developers, and educational institutions with access to OpenAI's powerful AI models, despite the recent decision by OpenAI to restrict direct API access from unsupported regions like mainland China and Hong Kong. This distinction highlights how Microsoft is stepping in to fill the gap, ensuring that AI in Hong Kong remains accessible for eligible users.

Azure OpenAI vs OpenAI: A Strategic Contrast

While OpenAI itself has blocked API access in Hong Kong, Microsoft’s Azure platform remains a critical gateway for users in the region to tap into OpenAI’s advanced AI models, including GPT-3.5, ChatGPT, Codex, and DALL-E 2. This service is accessible to eligible customers through Azure’s global infrastructure, ensuring continuity for AI-driven innovation in Hong Kong AI applications.

Key Points

• Continued Access: Microsoft has assured that there will be no changes to its Azure OpenAI Services for users in Hong Kong. AI models hosted in regions outside Hong Kong are still available, making it possible for developers to continue leveraging these tools for their projects.

• Microsoft's Advantage: While OpenAI has restricted its services in Hong Kong, Microsoft’s Azure maintains operational AI capabilities in the region. This distinction makes Azure OpenAI Services an essential resource for enterprises and educators needing reliable access to AI technologies.

• Enterprise Support: Launched in March 2023, Azure OpenAI enterprise support allows businesses in Hong Kong to apply AI models for commercial use cases, benefiting from enhanced compliance and security features tailored for enterprise environments.

Differences Between Azure OpenAI and Direct OpenAI Access

Although Azure OpenAI provides access to similar models as OpenAI, there are some notable differences:

1. Security and Compliance: Azure OpenAI offers enhanced security measures, ensuring that sensitive data is protected with enterprise-grade compliance. This makes Azure a preferred choice for businesses concerned with data privacy.

2. Integration with Azure Services: Azure OpenAI integrates seamlessly with other Microsoft Azure offerings, enabling users to create more customized, scalable AI solutions that are tailored to their needs.

3. Target Audience: While OpenAI's API is generally accessible to developers worldwide, Azure OpenAI is focused on providing AI models to businesses and educational institutions, with specific support for enterprise use cases in Hong Kong.

OpenAI’s Latest Models: Enhancing AI Reasoning with o1-preview and o1-mini

OpenAI has recently introduced two innovative models, o1-preview and o1-mini, aimed at advancing reasoning capabilities in artificial intelligence. These models offer significant upgrades in performance, particularly in complex fields such as mathematics, coding, and scientific research.

Overview of OpenAI o1 Models

o1-preview

• Purpose: Designed for complex reasoning tasks, o1-preview excels in solving intricate problems with a thought process that mirrors human reasoning. It's particularly useful in mathematics, coding, and scientific research.
• Performance: Demonstrating superior capabilities, the model achieved an 83% accuracy rate on International Mathematics Olympiad problems, compared to just 13% for GPT-4.
• Limitations: While powerful, o1-preview does not yet support features like web browsing or file uploads, which are available in older models like GPT-4o.

o1-mini

• Purpose: A smaller, faster, and cost-effective version of o1-preview, o1-mini is optimized for environments where speed and efficiency are critical, particularly in coding and STEM tasks.
• Performance: While o1-mini is 80% cheaper, it still offers strong performance in coding and reasoning tasks.
• Limitations: It lacks broad world knowledge and is most effective when all necessary context is provided within the prompt.

Both models are currently available for ChatGPT Plus and Team users, with plans to extend o1-mini to the free tier which would mean if you can find a workaround using a VPN then you should be able to access this model, although no specific date has been announced. Pricing for developers using the API starts at $15 per million tokens for o1-preview.

Accessing ChatGPT in Hong Kong: Workarounds and Alternatives

With OpenAI still restricting access to ChatGPT in Hong Kong, users face challenges when trying to leverage these new models. However, several workarounds are available:

1. Using a VPN
   A reliable VPN service is a common method to bypass geographical restrictions. VPN providers like ExpressVPN, NordVPN, and Surfshark allow users to connect to servers in countries where ChatGPT is accessible, such as the US or UK.

2. Third-Party Platforms
   Some platforms provide access to ChatGPT without the need for a VPN. For example, Studio Global offers services for Hong Kong users, enabling them to generate up to 22,000 words per month for free.

3. Microsoft Azure
   Microsoft Azure continues to offer OpenAI’s models for eligible customers in Hong Kong, providing another route for businesses to access these tools.

4. Local Alternatives
   The Hong Kong government has begun developing its own AI tools, potentially offering local solutions in the near future. While these are still under development, they could serve as viable alternatives to ChatGPT.

Conclusion

Despite the restrictions imposed by OpenAI, Azure OpenAI Services remain a reliable source of AI tools for users in Hong Kong. This continued access underscores Microsoft’s commitment to supporting AI adoption in Hong Kong, empowering businesses, schools, and developers to leverage advanced models like GPT-4 and Codex through the Azure platform. By offering robust security, integration options, and targeted support, Azure OpenAI is uniquely positioned to drive innovation and growth in Hong Kong’s AI ecosystem.

Although accessing ChatGPT in Hong Kong remains restricted, new AI models like o1-preview and o1-mini demonstrate the continual evolution of AI tools and the global buzz about their capabilities. For those looking to overcome these barriers, VPNs, third-party platforms, and services like Azure provide effective workarounds. As local alternatives emerge, Hong Kong may soon have more options for leveraging advanced AI tools.

For more detailed information on chatgpt, you can read additional articles on the South China Morning Post here and here.
If you do see the message, OpenAI’s services are not available in your country, you can find community support here. (error=unsupported_country)