On January 28, 2025, DeepSeek became the most popular AI term searched on Google for 24 hours, dethroning ChatGPT. This explosive record was achieved by the Chinese Artificial Intelligence assistant chatbot eight days after its launch. During the 24-hour takeover, the stock market experienced significant shifts in shares of major companies in the AI sector. Nvidia took the brunt, with its shares plummeting by 17% and losing almost $600 billion in market value.
By the end of January 2025, the DeepSeek AI assistant app became the most downloaded free application on the Apple Store in the U.S. This Chinese AI was developed by QuantumThink Labs in Shenzhen, China. It debuted in late 2023 and is giving OpenAI and Meta AI a run for their money. In this article, we discuss exactly why DeepSeek has disrupted the AI market and the challenges it is facing in its rise.
Why is Deepseek the center of attention?
Between 2021 and 2023, the world experienced an AI frenzy with the rise of Large Language Models and the massive adoption of AI assistant tools such as ChatGPT. It is thanks to this foundation that new entrants into the AI space, like DeepSeek, are disrupting and reshaping the norm in this AI-driven era. The two main shifts that have made Deepseek the center of attention are:
- Competitive solutions
Today’s generation turns to the internet for everything, both simple and complex. From relying on casino review sites to know where to play for the most advantageous bonuses to utilizing advanced tools for generating comprehensive monthly business revenue reports. AI has proven to provide a new layer of expertise, and DeepSeek has taken this feature to a new level.
DeepSeek’s flagship AI model, R1, has proven to excel in natural language understanding, mathematics, and coding. Its performance rivals similar models from OpenAI, Meta, and Google. It currently outperforms all other AI models in linguistic proficiency, especially in languages such as Chinese, Swahili, and Arabic. This wide proficiency has made the utilization of the AI model easy for millions of global users.
- Cost efficiency
DeepSeek claimed that the cost of training the R1AI model was approximately $6 million. This figure is a fraction of the cost invested in developing the popular AI model GPT-4, whose training investment stands at almost $80 million. QuantumThink Labs has successfully challenged the notion that advanced AI development requires bottomless funding. This new reality has caused a lot of key players in AI to rethink their development strategies and encouraged smaller AI developers that with smart engineering and strategic resource allocation, they can develop successful AI models.
Despite the very successful launch of DeepSeek, the Chinese AI model has not been immune to the challenges of privacy concerns that have plagued this new era of AI. These privacy concerns mainly revolve around data collection, cybersecurity, AI models, and governance. In the wake of its launch, several governments have banned or restricted its use in their countries.
Governments that have banned the use of Deepseek
DeepSeek requires users to create an account to use the AI model. This requirement has raised many alarm bells regarding data privacy, and the following countries have prohibited its use fully or to a certain extent.
- Italy: The Italian government was one of the first countries to ban the use of DeepSeek. The authority requested information confirming the personal data collected, the sources, the purpose of collecting the data, the legal basis of processing, and the location of the servers storing the data. The authorities cited the lack of this information as a potential risk of data breach.
- South Korea: The country’s defence ministry has blocked DeepSeek from accessing its internet-connected military computers. This move was taken as the country’s Personal Information Protection Commission (PIPC) awaits QuantumThink Labs to provide clarity on how it manages information of users using DeepSeek.
- India: The country’s central government has prohibited the use of AI tools such as DeepSeek on office computers and devices by all employees to minimize risks to the confidentiality of government data and documents.
- Australia: Acting on advice from its security agencies, the country has banned the use of DeepSeek on all government devices, citing potential malware threats and privacy risks.
- US: In the country, Texas state has banned the use of DeepSeek while the American Navy has restricted the use of DeepSeek for any work-related tasks.
- Taiwan: The country’s Ministry of Digital Affairs (MODA) officially communicated that potential data leaks from the AI model could expose sensitive data to Beijing, resulting in national security risk. MODA prohibited the use of DeepSeek by government employees in central and local government agencies, as well as in public schools, state-owned enterprises, government-funded foundations, and other government-affiliated institutions.
While these restrictions are a setback for DeepSeek, they do not come as a surprise. The Chinese AI-powered model has successfully cemented its position in the AI race to shape the future.
