Comparing Perplexity AI and ChatGPT: A Faculty Perspective

In the dynamic world of artificial intelligence, Perplexity AI and ChatGPT are two standout tools transforming how we teach, research, and collaborate. Each offers unique strengths and some limitations. Here’s a brief look at some of their differences and how they can enhance academic life.  These observations are based on our experiences using these two AI engines and some explicit comparison of responses to the same prompts by each AI engine.

Accuracy and Reliability: Stay Critical

Neither Perplexity AI nor ChatGPT are infallible, and both can produce incorrect answers. It’s crucial for users to critically evaluate the responses from these AI tools. While both models are trained on vast datasets, their ability to provide accurate and contextually appropriate answers can vary. Very often, ChatGPT needs to be poked multiple times to route to the correct answer. It’s vital to approach their responses critically and cross-check with reliable sources. Perplexity can also incorrectly interpret information it gleans from the internet, providing responses that are incorrect or incomplete.  Being able to recognize these potential errors is an important skill when using AI.

Sharing Capabilities: Perplexity’s Unique Link Feature

One distinctive feature of Perplexity AI is its ability to generate unique links for sharing responses. This can be particularly useful for collaborative projects, allowing users to easily share AI-generated insights with colleagues or students. ChatGPT, while powerful in generating detailed and customized responses, lacks this specific sharing functionality, potentially limiting the ease of collaboration.

Web and Video Search: Perplexity’s Edge

Both AI tools can search the web for information, but Perplexity AI goes a step further by attempting to parse YouTube videos for relevant content. This can be advantageous for users looking for multimedia resources or video-based explanations. You can also upload images to Perplexity AI for identification, interpretation, and caption creation.  ChatGPT, on the other hand, focuses primarily on text-based web content and does not currently include video search capabilities.

Code Generation and Calculations: ChatGPT’s Strength

For faculty and students in fields that require computational work, ChatGPT stands out with its ability to generate and execute Python code. This functionality allows users to perform calculations, create graphs, and even download the generated code for further use. Although Perplexity AI can generate code, it does not offer this level of integration with coding tasks, making ChatGPT a better choice for disciplines where computational tools are essential.

Discipline-Specific Expertise: ChatGPT’s Depth in Biochemistry

In discipline-specific inquiries, such as biochemistry, ChatGPT demonstrates superior capabilities. For example, ChatGPT can calculate the isoelectric point (pI) of a peptide by utilizing known pKa values, a task where Perplexity AI falls short, often admitting a lack of specific data. This depth of knowledge makes ChatGPT a more reliable tool for specialized academic queries.

Sign-In Comparisons and User Experience

Both Perplexity AI and ChatGPT provide free options without an account, free options with an account, and paid options. Without logging in, Perplexity AI tends to offer a better experience. When logged in or using paid versions, the experiences and features of both platforms can vary, with ChatGPT offering enhanced computational and visual data generation capabilities.

Creating Figures from Data: ChatGPT’s Performance

When it comes to generating figures and visual data representations, ChatGPT outperforms Perplexity AI. While ChatGPT still has large room for improvement, with persistent tweaking, it can produce acceptable data illustrations, which can be valuable for researchers and educators. Perplexity, in its paid Pro version, has recently integrated visualization using other AI engines.

Conclusion

Both Perplexity AI and ChatGPT offer valuable features that can support academic work. While Perplexity AI excels in multimedia search and sharing capabilities, ChatGPT provides superior computational and discipline-specific support, for example in biochemistry. Faculty and students are encouraged to leverage the strengths of each tool, critically assess AI-generated content, and choose the platform that best meets their academic needs.

For more insights and updates on utilizing AI in academia, stay tuned to the Office of Faculty Excellence blog.

Disclaimer: This blog post was generated by ChatGPT-4o based on key points provided by faculty and then reviewed and edited for accuracy. The experiments were performed in June 2024. As AI is rapidly evolving, the results may not reflect future capabilities.