By Piers Hamilton, Fourth Year, French and German
ChatGPT, introduced by OpenAI in late 2022, was a key milestone in bringing artificial intelligence to the public. The free AI Chatbot processes language and commands at revolutionary speeds in over 80 languages, spanning from English all the way to Faroese. The international community has widely adopted OpenAI, with educational institutions and major corporations incorporating its AI technology into research, automation, and product development. A prime example is Apple’s integration of ChatGPT technology in the new iPhone 16, bringing ‘Apple Intelligence’ to life.
AI’s expansion hasn’t been met with unanimous approval - many universities and members of the public have become increasingly hesitant of its growing influence and use. OpenAI’s ChatGPT has been banned in China due to its ability to generate and disseminate information that conflicts with the country’s strict censorship laws and AI regulations. For similar reasons, North Korea, Russia, and Iran have also imposed bans, citing concerns over AI’s potential to challenge their own systematic frameworks.
Backed into a corner with these restrictions, Apple had little choice but to integrate its Apple Intelligence with Chinese owned Alibaba’s Qwen AI software for the Chinese market.
Given the polarised nature of today’s world, political, economic, and technological rivalries are increasingly taking centre stage. Just as during the Cold War, when the Soviet Union responded to Coca-Cola’s ban with its own alternative, Baikal, we are now witnessing a similar development in the realm of artificial intelligence: the emergence of DeepSeek.
DeepSeek, a Chinese AI Large Language Model developed in 2023 by Hangzhou-based Liang Wengfeng, mirrors many of the functionalities of ChatGPT. It operates similarly to OpenAI’s software, with users able to type commands into a textbox to receive near-instantaneous AI-generated responses. Capable of generating and refining text, solving complex mathematical equations, and writing code, DeepSeek has become a steadfast adversary in the world of AI.
While many experts have argued that DeepSeek’s capabilities fall short of ChatGPT’s, it is impossible to overlook the remarkable achievement of Wengfeng and his team. – a reported $6 million AI training cost in comparison to OpenAI’s expenses that were supposedly ‘over $100 million’ in the development of GPT-4.
On top of this, DeepSeek could surpass ChatGPT in the emerging discussion on AI’s environmental footprint. With fewer chips needed and improved energy efficiency, DeepSeek is positioning itself as a more sustainable option compared to the AI systems from Silicon Valley’s tech giants.
So what does this all mean? Have we entered an AI Cold War or is the market simply diversifying?
Maria Zhang, an associate professor at the University of Sydney stated that DeepSeek is a display of China’s ‘growing technological prowess, with its challenge its challenge to the Western-dominated sector is pushing for greater domestic innovation, promoting self-sufficiency, and fuelling the rise of ‘tech isolationism’.
DeepSeek has rapidly become a fixture in the daily lives of many Chinese people. According to a recent BBC report, the software is now being used instead of traditional therapy, with interviewees claiming that DeepSeek offers free counselling services that surpass any paid alternatives they've experienced.
But can DeepSeek rival ChatGPT in its appeal to the student community? It’s a complex and multifaceted question. One key factor shaping the Chinese alternative is China’s censorship laws. For example, topics like the Tiananmen Square events won’t generate any output, which is a significant limitation to consider when using the software. While DeepSeek remains current with up-to-date articles and offers an extensive reach, it’s often seen as less versatile in specialized subjects compared to ChatGPT.
Can DeepSeek truly rival ChatGPT in its appeal to the student community? It’s not a straightforward answer and by taking a step back has multiple layers to it. One significant limitation is China’s censorship laws that dictate much of DeepSeek’s operations. The software is restricted from discussing sensitive topics like Tiananmen Square, which could impact its utility for students seeking unbiased and comprehensive information. In addition to this, while DeepSeek is generally up to date with articles and offers an extensive range of sources, its strength in versatility and general knowledge comes at the expense of depth, often making it less reliable than ChatGPT for academic or niche subjects.
That said, DeepSeek operates fundamentally differently from ChatGPT. Unlike ChatGPT, which often restricts users behind a paywall for its premium service, DeepSeek remains completely free to use. Additionally, it has earned praise for its impressive mathematical and coding capabilities, achieving a remarkable 90% accuracy on highly complex problems.
As AI-powered plagiarism detectors become an ever-growing fixture in university grading systems, DeepSeek may quietly emerge as a more appealing option for students looking to sidestep detection. While systems designed to flag ChatGPT generated content focus on identifying repetitive phrasing and word patterns, DeepSeek’s approach to language is distinct enough to potentially slip under the radar. Despite its limitations in creativity, could the Chinese AI prove to be a more elusive tool for those seeking to avoid the growing scrutiny of AI generated work.
Ultimately, the world of Artificial Intelligence is evolving at a rapid and unpredictable pace. Language models like ChatGPT and DeepSeek are not only reshaping the way humanity progresses with AI, but also amplifying the divisions arising in the 21st century. As these technologies become more accessible, they become increasingly empowered in redefining how we interact with information, and how those very interactions may drive us further apart.
