In November 2022, an artificial intelligence (AI) and research company, OpenAI, launched a free AI powered chatbot service called ChatGPT.
ChatGPT is a language processing tool that can answer questions and assist users with completing tasks, such as composing emails or writing essays. The service is free to the public and currently has over 100 million users, according to statistics published by Exploding Topics.
AI and ChatGPT has been trained by processing a variety of information on the internet, including websites, books, news articles, blogs, and more. Once absorbing the data out there, ChatGPT programmers used Reinforcement Learning from Human Feedback (RLHF) to create natural human conversational dialogue. By using RLH, AI trainers provided the model with conversations in which they played both the user and AI assistant, according to OpenAI.
Since the launch of ChatGPT, the quickly growing software has stirred up mixed responses from the public. Some companies, such as Microsoft, encourage the use of AI powered tools to increase efficiency in its programs.
JaQualon Taylor, a systems engineer at Medix Dental IT, explains that AI can be a double-edged sword in a variety of situations. With technology being able to complete human tasks at a cheaper price, some employees worry that AI might threaten their careers.
“Currently, employees are worried that AI will take away jobs because of its capabilities,” explains Taylor, “however, using ChatGPT will allow for better productivity and a great chance for growth in a business.”
Taylor also notes the pros and cons of AI and ChatGPT in an academic setting. While many schools are banning the use of AI powered chatbots out of concerns that students will misuse the software and compromise their academic integrity, Taylor believes ChatGPT can be a great resource for students.
“ChatGPT allows students to research and find more information on what they’re needing to use,” says Taylor, “but just like alcohol, use it responsibly.”
ChatGPT also has its fair share of limitations. Because the software essentially learns from humans, the program can spit-out harmful or biased responses that it learned from humans. The AI software studies patterns to form responses and can occasionally produce irrelevant responses if ChatGPT does not understand what the user is asking.