Chat about ChatGPT: Will AI steal my job and do it better?

Being a first-generation immigrant, I always felt insecure about my job. So much that I almost look forward to be laid off. I was scared that my work visa would be denied, that my bosses finally find out I was nothing like the person I demonstrated during interview, that companies join the recent mass layoff and realize business can run just OK without me. I was never scared, however, of robot taking over and doing what I am doing. I mean, those are only in the movies, right? Furthermore, I work as a solution lead in the technology department, which I consider not simple enough that a robot can just take the order and fulfil.

Then last week I was introduced to ChatGPT from a YouTube video, and was instantly surprised by how much AI has advanced, to not only capable of executing orders, but also communicate it in such a natural and fluid way. I tried a few work related questions and was impressed by how much it knows about my company and its applications, about not only to write codes for a requirement, but also to explain it in layman’s term that I could understand, and of course, about how to ask for a pay raise. Here are a few findings from 2 weeks playing with ChatGPT. Hint: Not all findings are positive.

1.??????The good: It has the potential to 80% replace Google search. I say 80% only because I’m a conservative Chinese (not all Chinese are conservative, but I am) and don’t want to open doors for argument. I think it is 100%, at least for the searches I perform on Google. Occasionally I spent time searching things like how to use a specific MS Excel formula. Usually I ended up landing on a 20 minute YouTube video, impatiently forwarding and reversing to find the clip that’s related to the specific topic I’m interested, and watching at least one commercial of McDonald’s coffee in between. These two weeks since I accessed ChatGPT, I found myself Goggling significantly less. ChatGPT has no commercials or paid content, get right into the answer (usually in structured bullet points), and can re-generate the answer when I need.

2.??????The double edge sword: It not only gives search result, but also educates the person who performs searches. Go to any coding language forum, you will see people asking all types of specific questions because online search is only teaching the knowledge, not the actual solution to a particular problem. ChatGPT, however, gives the solution the specific problems. For example, we are currently working on some user data tracking and analytics requirements. I put in the formular to calculate bounce rates, and asked ChatGPT to write me codes in Python. After displaying the codes, it also explains the logic in writing them. Also, when I’m in middle of building a training site for a complicated application catering a large number of trainees, I asked ChatGPT things like “what’s a good training site structure”. The answers I collected have really helped me in sketching a clear skeleton and to build content. So far so good, right? However I think the other edge of the sword is that it took away time we spent in struggling – which I considered the best way of learning something. I remembered the correct answers the best, when I spent so much time thinking, getting the wrong answer, analyzing what went wrong, and repeat all over again. Using ChatGPT has taken away this opportunity to acquire something new, since the answer is one search away.

3.??????The bad: I know sometimes the answer that ChatGPT gave is wrong. I’m less concerned about that. Because when we land on an answer after significant time in Googling and researching, sometimes it is still wrong. What I don’t like, however, is that it sounds very convincing while giving the wrong answer. ?When I’m not very confident about my answers, at least I won’t look that assured. However, no one will be able to read between the AI lines and tell the likelihood of the answer being correct. One co-worker told me that she asked ChatGPT to summarize the content of a book and generated the answer twice, both times the summarization were completely off. ?

4.??????The Worry: I think by now the word has came to your mind already – Cheating. What’s the definition of cheating in workplace? Many years back I saw a piece of news saying a star performer in a company was found “outsourcing” his own work to a group of developers in Asia. Now that is clearly cheating. However, if I don’t know how to complete an analysis, and Googled the steps, would that be called cheating? If I wanted to identify stakeholders of an application in my company, and took the names from an earlier presentation, would that be cheating? Then instead of Googling and sourcing from older materials, if I got the answer from ChatGPT, would that be considered cheating? I wasn’t so worried about cheating, I was more worried that with tools like this, AI will become smarter and human will become dumber – then eventually people who has the habit of relying on AI to do their jobs will no longer be needed in their position – it was all done by AI after all, isn’t it?

5.??????The bottom line: As long as my company policy permits, I will user ChatGPT to improve my productivity at work. I will use it as a brainstorming partner to lay out idea and as a how-to guide to get quick answers. I will never directly copy and paste things just like I won’t do for searched results. More importantly, I want to get to know more about AI-related fields, industries and services. So when the time comes that our jobs are transformed by AI, I will have the new skills to adapt to this change.

To have a less predictable ending, I asked ChatGPT to write an article about “How surprised I am when finding ChatGPT” and added it below. Which article did you enjoy reading more, mine or ChatGPT’s? I’m curious to find out whether I gave some “personalities” to my article at least!