
DeepSeek R1: A New Leader in AI and the Questions of Fair Competition
In recent weeks, the Chinese artificial intelligence model DeepSeek R1 has garnered significant attention from businesses and the tech community. Just days after its release, it positioned itself as a powerful, affordable, and versatile alternative to Western AI models. But concerns have emerged…
Franc Smidt, Journalist
Germany, 27.01.2025
Unexpected, Affordable, and Powerful
DeepSeek R1 is an open-source artificial intelligence model designed for data analysis, logical reasoning, programming, and content creation. With 20 million hours of training on high-quality data and advanced algorithms, R1 demonstrates impressive efficiency at significantly lower costs. Companies like ZoomInfo have already started using it to replace more expensive AI models, reducing their AI-related expenses by two-thirds.
The model’s uniqueness lies not only in its cost-effectiveness but also in its open architecture. Companies can download and integrate it into their own cloud systems, avoiding risks associated with storing data outside their jurisdiction. This makes it particularly appealing for businesses looking to minimize costs while maintaining control over sensitive information.
One of R1’s key innovations is its ability to understand and process complex tasks while keeping operational costs low. For example, it is already being used to create data analysis tools and develop solutions in sales, allowing companies to offer more personalized and strategically refined services to their clients.
The arrival of DeepSeek R1 has not only caused a market shake-up but is also driving significant changes in the technology sector. Companies that have traditionally dominated the AI space, such as OpenAI and Anthropic, now face a new reality, where open AI models are becoming formidable competitors. However, new concerns have surfaced…
Microsoft and OpenAI Investigate Possible Unauthorized Use of Data
Leading technology giants have raised concerns that the DeepSeek AI platform may have used OpenAI’s data to train its models without proper authorization.
Reports indicate that Microsoft and OpenAI launched an investigation after security specialists detected large-scale data extraction linked to individuals associated with DeepSeek. If confirmed, this could violate OpenAI’s terms of service, which explicitly prohibit the use of their data for training competing AI models.
There are suspicions that DeepSeek may have employed a technique called “distillation”, where a smaller model is trained based on the outputs of a larger one. While this practice is common in the AI industry, using it without proper authorization could constitute an infringement of intellectual property rights.
An OpenAI representative stated that companies from various countries, including China, are constantly attempting to leverage the models of leading U.S. AI developers. In response, OpenAI is implementing protective measures and collaborating with the U.S. government to prevent such incidents.
The outcome of this investigation could have far-reaching implications for the relationship between Western and Chinese AI companies, as well as for the broader approach to intellectual property protection in the rapidly evolving AI sector.