An economics professor says he’s stunned by the progress made by ChatGPT, a chatbot developed by OpenAI and author Sam Altman earlier this year. Bryan Caplan of George Mason University was taken aback by how quickly the digital assistant had improved its performance on his midterm exam from a D grade in January to an A.
To test the GPT-4 update, Caplan posed the same questions from his December midterm in a blog post he wrote since then. While the crude ChatGPT-3.5 was unable to understand basic economic theory, much less provide accurate responses, the GPT-4 version stunned Caplan by offering concise and clear answers, being able to correctly explain and evaluate concepts proposed by economists such as Paul Krugman and achieving a perfect score.
At the same time, AI bots have also caused issues for examiners, as it can be difficult to detect if a student has plagiarized responses from ChatGPT. To prevent this, Caplan is considering eliminating graded homework in favor of constantly changing exam questions.
It’s clear that the AI revolution is well underway, with tools such as ChatGPT completely revolutionizing the face of economics education. As the AI industry continues to progress and become more complex, it will be intriguing to see how much further ChatGPT can go. Additionally, Professor Bryan Caplan’s bet with AI doomer Eliezer Yudkowsky about the end of the world prior to 2030 is another great example of how AI is affecting our lives and our future.
