Tommy Liu 531253tl, 1 October 2023
I was at the end of my first masters program in Marketing Management when Chatgpt gained a lot of popularity. A lot of courses were not able to adapt and it was very easy to use Chatgpt as a supportive tool to help answer quiz questions for some courses. However it was quick to realize that the outputs given out of Chatgpt, were not always correct. This started the process for myself to train, giving Chatgpt the correct inputs to gain the best outputs possible. Even Though it is artificial intelligence, the process felt like you were talking to a person who can only give you the right answers when asked the correct questions. During the months after the first usage, the knowledge of how to use Chatgpt became better, and I was able to adapt. Boundaries for Chatgpt were that it was not that good at giving opinions and if you ask something borderline illegal, it made sure it does not give answers that can break the law.
Besides using Chatgpt for education, I used the tool also for other different things. For example as a stock investor, I tried to use Chatgpt to maybe write a trading or investing strategy for me. However, quickly the flaws of Chatgpt came through and it is of course not possible to make-up a 100% profitable investing strategy. This is when I realized Chatgpt is more of a supportive tool than an active decision making tool.
Lastly there are also dangers with Chatgpt that I encountered. When starting my current internship I was working on some Excel documents. I was planning to insert a table with data, to ask if it could give me python prompts to run for valuable analysis. However, I discussed this with a colleague, explaining why this is not allowed in the company. Chatgpt gets trained by all the inputs it gathers, if you put company specific data into the program, it is able to use this data for other users. Which is ofcourse not what a big corporation wants.
All in all, I encountered a lot of situations with Chatgpt and learned the useful parts and the parts that you have to be careful of. Use Chatgpt, but use it supportive and wisely!
Interesting to read how you made use of Chat-GPT in your past. It is also valuable to understand that Chat-GPT should be used with care. AI is very useful if used correctly. As you say, people should always stay in control. Therefore, it is very important that people get the right education about this tool. This could be relevant at universities (maybe include it in courses even?), but also inside big companies as you mention. This could make sure that users of generative AI tools understand the potential flaws of the tool instead of being so enthusiastic and only focusing on the positive points.