Is DeepSeek AI A Game-Changer Or A Risky Bet?

The release of DeepSeek AI’s R1 model has sent shockwaves through the tech industry, leading to the biggest single-day stock decline in U.S. history for microchip giant NVIDIA, which lost nearly $600 billion in value. Founded by Chinese entrepreneur Liang Wenfeng, DeepSeek AI model is reportedly cheaper and faster to develop than competitors like ChatGPT and Gemini AI. While initial reports suggested a cost of just $6 million and two months for development, independent research firm SemiAnalysis estimates the actual figure to be around $1.3 billion.

As an open-source AI model, DeepSeek is free to use, prompting a surge in downloads from marketers and tech professionals alike. But is it truly a better alternative to existing AI platforms?

How DeepSeek Compares

James Heimers, senior vice president of analytics at creative agency Rapp, notes that comparing DeepSeek to ChatGPT isn’t straightforward. Instead, it should be assessed alongside both proprietary models like Gemini and Claude and open-source models such as LLaMA and BLOOM.

What sets DeepSeek apart is its Mixture-of-Experts (MoE) design, which enhances efficiency by activating only the necessary parts of the model for specific tasks.

Michael Ruby, president and chief creative officer of B2B agency Park and Battery, highlights DeepSeek’s exceptional speed: “It’s faster than ChatGPT and other AI platforms we’ve tested, seemingly without sacrificing accuracy.” However, he argues that speed alone isn’t enough to distinguish DeepSeek, noting that it requires coding expertise to build AI agents, making it less accessible for marketers compared to platforms like ChatGPT.

Despite this, Esme Robinson, director of Platform Solutions at Epsilon, sees potential for marketers. She believes that as AI evolves into ‘agentive AI’—where generative and predictive AI combine to create real-time ads—DeepSeek could play a crucial role. However, she warns that human oversight remains essential to ensure content aligns with brand identity.

Dave Porter, global director of technology integration and innovation at media agency Oliver, sees great marketing applications for DeepSeek, from automated content creation and SEO optimization to sentiment analysis and personalized ad campaigns. He emphasizes its potential in data-driven marketing but acknowledges that security concerns may limit its adoption.

Data Security Concerns

DeepSeek’s Chinese origins have raised red flags in the West, particularly regarding data privacy. DeepSeek stores data on Chinese servers, making it subject to China’s strict data laws, which could allow government access. Additionally, reports linking DeepSeek to China Mobile—a company flagged by the U.S. for ties to the Chinese military—have heightened concerns.

In response, the Italian Data Protection Authority has already banned the AI bot, while Australia has prohibited its use on government devices. The U.S. and other nations are also considering similar restrictions.

The Road Ahead

Despite these concerns, DeepSeek should not be dismissed outright. It’s too early to determine its long-term role in marketing but acknowledges that generative AI is becoming an indispensable tool. If DeepSeek can address privacy issues, it may earn a place in marketing AI stacks.

The future is owned by businesses that not only utilize AI but also manage it within their marketing strategies. The competition among AI models is far from finished.