Entrepreneurial Showdown: ChatGPT vs. Wharton MBAs

0:00So Christian Terwish, who is a professor at the Wharton School recently unleashed a groundbreaking experiment involving, of course, Chad Chippity. And the experiment aimed to essentially guess whether Chad Chippity could outperform MBA students in really coming up with new product ideas. So after successfully passing an MBA final exam earlier this year, Chad Chippity is now facing its next challenge. Is it actually going to be better, faster, cheaper for creativity and especially in entrepreneurship and coming up with new businesses? So Terwish expressed his astonishment at the experiment's results in an interview with Wharton Business Daily and said, I was really blown away by the quality of the results.

0:43I had naively believed that creative work could be the last area in which we humans would be superior at solving problems. So the details of this kind of AI versus human experiment have been published in a white paper co-authored by him, along with Wharton OID professor Carl Earl Rich, who's a former Wharton graduate student. And then we had Leonard McKinn and Cornell Tech professor Karen Girotra. So the initial setup began with 200 product ideas from MBA students, specifically designed for the college student market and priced under$50. These ideas, noteworthy for being created before the release of ChatGPT and similar generative AI tools, served as a baseline.

1:27and the research team then fed the same assignment into GPT-4, generating another 200 ideas, 100 with no initial context and another 100 seeded with examples of promising products. So completing the assignment in a mere 15 minutes, I'm sure this doesn't surprise anyone that is, you know, actively uses ChatGPT, but ChatGPT actually outpaced human creativity, which typically generates about five ideas in the same time frame. But the speed wasn't the only criteria. The team also assessed the visibility of the ideas through a blind consumer survey. I love this, right? Because of course, it's hard to say like which of these ideas would be the best if it's, you know, like the people running the test or if it's, you know, asking chat GPT or like a specific person, but, you know, running a survey and saying, you know, which of these do you think is better?

2:19Definitely is a good, you know, a blind consumer survey, great way to test it. So in this department as well, ChatGPT's ideas came out on top. The average purchase probability of a product idea from ChatGPT was 47 % compared to 40%, which was the rate for human generated ideas. And I think what else is really interesting here is that when it was like seeded with good examples, right? Like, so essentially, if you tell ChatGPT, and this is, I think this is really important because it's going to come back to something interesting. but when you give it like some good examples chat gpt's um preference raising rows to 49 so slightly higher but i've actually seen it go much higher in certain areas where i'm like hey i need x y and z especially in an area that chat gpt may not have a lot of training data on so i'm like hey i need xyz here's an example just like write me use this example and write me something but use all of these facts and make it in this specific area market niche whatever ChatGPT does a really good job of like, you know, kind of getting the concept and ideas out of like out of a main piece of content or main document and then, you know, elaborating or changing them to your specific needs.

3:27So Turwish, who also serves as the co-director of the Mac Institute for Innovation Management, highlighted three major insights from this whole endeavor. endeavor. First, he said chat GPT can be a robust tool for generating creative ideas. Obviously, I'm sure we all know that. And said, quote, it doesn't matter whether they are entering the local business plan competition or working on the cure for cancer. It's cheap, it's fast, it's good. What's not to be liked? Worst case is you reject all of the ideas and run with your own, but your research, but our research speaks strongly to the fact that your idea pool will get better.

4:02So, you know, I think he's really kind of leaning on the fact that you can use this to help spark ideas and innovation. And I think that's 100 % true. A lot of times, though, like we saw in the study, it just was straight up better. Like, if you seeded it, it was almost 10 % better than humans. And if it was just going straight up, it was still 7 % better, which is quite significant. So second, he believes that generative AI tools can accelerate the pace of innovation by not only generating ideas, but also aiding in their evaluation. So after all, you know, kind of sorting through a sea of ideas or, you know, trying to find the, you know, a needle in a haystack, right, is essentially where human expertise is very indispensable.

4:46But AI is really good at kind of accelerating everything by creating stuff. So the last thing that he noted is that he really kind of emphasized that the conversation shouldn't be framed as a competition between man and AI, but he said, quote, we need to find a way in which AI becomes your creative co-pilot together. You can become a more innovative team. Um, now this is, I've talked about this a lot, so I won't give you my whole spiel, but I really do think there's a lot of areas where an AI, especially as it's continues to get better and better will like completely take over in, even in creativity and a lot of different areas, we'll be able to make a lot of different decisions, um, and do a lot of things.

5:23Now, I think, you know, a lot of people may disagree with me on that. But regardless of whether you disagree with me on it, the fact that it's faster and cheaper means inevitably companies and organizations will implement that. So whether you like it or not, it's going to happen. And so I guess at this point, we kind of have to ask the question of like, if an AI is running things or making decisions or, you know, creating content, like, do we need to regulate it? How do we need to, you know, address it? Do we need to have disclaimers on this kind of content? Like you can have all those kinds of conversations.

5:54There's no way, you know at this point we can say an ai won't or shouldn't or can't do the thing because it definitely can and it definitely will especially because it's cheaper and faster and is you know this case literally showed that it was better right like it came up with better ideas and so i think um you know humans augmented with this of course are going to outperform other humans that do not have it but in a lot of places like what happens when the ai outperforms all humans no matter what, right? So very, very interesting conversations. I think as AI continues to advance, the challenge may not be about whether it can replace humans creatively, but rather, you know, some people say it's not about that, but it's kind of about how to integrate the tools into our workflows and kind of augment human potential.

6:37I think that's conversation. But again, I really do believe AI will replace a lot of what humans do. And if you don't like the output, it's just a matter of time before the models get advanced enough and the prompts get advanced enough and fine-tuned enough that these things are able to outperform humans on a majority of tasks. I think it's coming. So of course, that's my opinion, but I'd love to hear what you think. Make sure to let me know. Join our Facebook or Discord groups and let me know your opinion.