AI unveiled: a comparative analysis of code accuracy and reliability across various chatbot platforms
With the increasing number of AI chatbots being introduced to the market, they are now being utilised across various industries. AI chatbots can perform a wide range of tasks, including planning, designing, mathematics, engineering, and scientific programming. The surge in popularity of AI chatbots...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project |
| Language: | English |
| Published: |
Nanyang Technological University
2024
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/176352 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Nanyang Technological University |
| Language: | English |
| Summary: | With the increasing number of AI chatbots being introduced to the market, they are now being utilised across various industries. AI chatbots can perform a wide range of tasks, including planning, designing, mathematics, engineering, and scientific programming. The surge in popularity of AI chatbots began with the introduction of ChatGPT in November 2022. Subsequently, major technology companies such as Microsoft with Bing AI and Google with Google Bard followed suit, further expanding the landscape of AI chatbots. These AI chatbots have evolved to possess enhanced capabilities, such as image generation and puzzle solving.
Due to the significant dependence on these chatbots, ensuring the reliability and accuracy of the solutions they provide is crucial. This project aims to investigate whether AI chatbots can deliver accurate and dependable programming code for engineering projects. The chosen engineering project for this study is an automatic drone inventory counting system. The objective is to generate code solely from AI chatbots and execute it to enable the drone to fly to a designated area, utilise its camera to scan products, and automatically record the inventory counts in an Excel sheet.
To ensure the accuracy of the AI-generated code, the project will be divided into stages. Consistent starting input prompts will be used for all tested AI chatbots to ensure fairness in evaluation. Initially, a selection of popular and freely available AI chatbots was made, followed by a series of preliminary tests to narrow down the choices to three. Subsequently, the selected AI chatbots will undergo testing in stages, starting with object recognition and progressing to inventory counting. The tests will be recorded and evaluated based on criteria such as accuracy and completeness of the results.
Recommendations for future research include investigating the impact of input prompt quality on the response quality. Additionally, utilizing a drone equipped with enhanced processing capabilities can ensure the accurate and immediate recording of captured frames. Another avenue for future research involves conducting experiments on premium AI chatbots that require subscription or payment to determine if their responses differ in quality from those provided by free chatbots. Exploring these recommendations can further validate the reliability and quality of answers provided by AI chatbots. |
|---|