Introducing Code Interpreter: A Potential Game Changer for Education, but with Limitations for Bioinformatics
Researchers at West Virginia University have identified the potential of the new ChatGPT plugin, Code Interpreter, in educational settings. However, they have found limitations when it comes to its use in the field of bioinformatics, which involves computational methods to prioritize targeted treatment for cancer and genetic disorders.
Gangqing “Michael” Hu, assistant professor in the Department of Microbiology, Immunology and Cell Biology at WVU School of Medicine, acknowledges the benefits of Code Interpreter for coding in STEM fields in an educational setting. However, he points out that the plugin lacks the necessary features for bioinformatics. Nonetheless, Hu believes that future developments of Code Interpreter might address these technical issues and expand its usability to fields like bioinformatics, finance, and economics.
The Background: ChatGPT and the Need for Code Interpreter in Biomedical Research
ChatGPT, an artificial intelligence chatbot, has gained popularity since its release in December 2022. However, it has not met the needs of those working in biomedical research, particularly bioinformatics. Scientists were hopeful that OpenAI’s Code Interpreter plugin would bridge this gap.
Hu and his team conducted tests to evaluate Code Interpreter’s features. Their findings, published in the Annals of Biomedical Engineering, show that while the plugin does improve access to coding for those without a science background, it still has limitations. Users need to interpret data, ensure the accuracy of results, and know how to interact with the chatbot.
Bioinformaticians heavily rely on coding, computer software programs, and internet access to store, analyze, and interpret biological data. Despite Code Interpreter’s advantages, including minimizing fictitious answers (hallucinations), Hu acknowledges specific improvements needed for bioinformatics.
The Limitations and Potential of Code Interpreter in Bioinformatics
While Code Interpreter has its advantages, it has certain limitations in the field of bioinformatics. The plugin currently only supports Python as a programming language and lacks software packages dedicated to bioinformatics. It also does not provide access to internet data and is limited in handling large files and parallel processing.
However, Hu believes that Code Interpreter has the potential to improve. He suggests upgrades such as internet access, software installation for bioinformatics, expanded storage capacity, and support for additional programming languages. Privacy and security features are also necessary to comply with regulations like HIPAA.
Despite these limitations, Hu plans to introduce Code Interpreter to his students for data visualization in future classes. He hopes that future upgrades will address the limitations and enable broader use of the plugin in bioinformatics coding.
In conclusion, while Code Interpreter shows promise in educational settings, it still needs improvements to cater to the specific needs of bioinformatics. Hu remains optimistic and continues to monitor new AI programming and features for future applications and innovations.