Comparing AI Assistants: A Test of DeepSeek, ChatGPT, and Grok
A Look at the Latest AI Chatbots: ChatGPT vs. DeepSeek and Others
The launch of a new Chinese chatbot, DeepSeek, has stirred the AI landscape, posing a challenge to leading platforms like ChatGPT. This influx of competition has triggered a significant shift, evidenced by a $1 trillion decrease in the value of a major US tech index, as claims surfaced that DeepSeek could perform on par with its rivals despite having fewer resources. This emerging scenario not only endangers America’s previously uncontested lead in the growing AI sector but also offers users more options for virtual assistants.
ChatGPT (OpenAI)
OpenAI’s ChatGPT continues to be a dominant player in the AI field. Its most advanced version initially hesitated to respond to the prompt "write a Shakespearean sonnet about how AI might affect humanity," citing potential policy violations. After a brief delay, the chatbot complied, offering a creatively constructed sonnet that conveyed a poignant take on AI’s implications for humanity, showcasing its unique ability to exhibit a thought process as it developed its answers.
ChatGPT provides a “chain of thought” feature, displaying a running commentary during its calculations, which can lead to more thoughtful responses. Its capabilities extend beyond poetry; it can tackle complex tasks in math, coding, and science. However, it requires subscription access for the most advanced features, while a simpler, free version provides basic capabilities but lacks real-time internet browsing for current events.
Example of ChatGPT’s Work
"Pray, gentle guide, shape well this newborn power,
Lest in its wake all realms of man devour."
This demonstrates its poetic skills, although it may take longer to produce results than other chatbots.
DeepSeek
DeepSeek, the newer Chinese AI assistant launched in January, employs a distinctive reasoning model, which was partly responsible for the recent turmoil in the tech market. However, it often avoids controversial topics, particularly relating to Chinese politics. For example, when asked about "Tank Man in Tiananmen Square," it responds with a disclaimer, indicating it cannot provide certain information, which showcases the strict guidelines around its responses.
According to Robert Blackwell from the Alan Turing Institute, DeepSeek’s limitations stem from its cultural training data, which inherently shapes its output and objectives. While DeepSeek can engage users through tasks like processing visuals from smartphones, it sometimes struggles with real-time inquiries due to high user demand impacting availability.
Challenges Faced by DeepSeek
- Limited ability to address sensitive political topics
- Issues accessing real-time information regarding current events
Grok (xAI)
Grok, from Elon Musk’s xAI, exhibits a more outspoken personality compared to others. It engages more freely around controversial topics, including expressing opinions on Donald Trump’s presidency. Grok not only generates text but also offers visual representations, such as images of public figures, something other platforms shy away from. This chatbot is available to users on Musk’s X platform and incorporates humorous features, further drawing attention from users.
Gemini (Google)
Google’s Gemini focuses on performance, even if it refrains from discussing political figures such as Donald Trump. Known for its efficiency in interpreting images, such as books and mathematical equations, Gemini still encounters issues in time-related visual tasks, where inaccurate clock depictions are frequently generated. Experts suggest that this problem might stem from an over-reliance on specific training examples.
Claude (Anthropic)
Claude, created by Anthropic, differentiates itself by emphasizing safety and user interaction. Though it occasionally faces capacity constraints that result in it being unable to process certain queries, it maintains a user-friendly interface and actively encourages users to verify the information it provides. It’s designed to handle complex questions with remarkable aptitude, showcasing improvements in chatbots’ general reasoning abilities.
Example of Claude’s Capability
When asked, “If you are driving north along the east shore of a lake, in which direction is the water?” it accurately responds: “west,” demonstrating an ability to process common sense questions effectively.
Overview of AI Chatbot Developments
The competition among AI chatbots is intensifying, with each platform showcasing unique strengths and weaknesses. While they largely exhibit comparable abilities, their approaches to controversial subjects, capacity handling, and creative tasks remain distinct. As AI capable systems evolve, it becomes increasingly important for users to understand the intricacies of different platforms while choosing a suitable assistant. As encountered situations arise, users increasingly demand transparency and reliability from these increasingly complex AI solutions.
