My Biggest 2022 Prediction: GPT-3 will take over schools and college campuses
My next substack article will be a major piece on my GPT-3/AI predictions for the 2022 new year.
However, there is one prediction which I’m really confident about and felt it’s worth exploring here in more detail on its own. I mentioned this towards the end of my last podcast:
Now that OpenAI’s GPT-3 is open to the public, I think we will start seeing many more students using GPT-3 and spreading it actively on college campuses.
This is a topic I’ve been following since the beginning of GPT-3. For students, GPT-3 can help them with any writing task such as writing essays, summarizing research papers, writing poetry, and if you went to business school, even generating SWOT analysis tables. It has ideation value too, such as helping you come up with an essay topic or thesis. GPT-3 really helps eliminate the problem of writer’s block as well, which I remember constantly battling as a university student. GPT-3 can play the role of a tutor or simply explain advanced ideas to students 24/7 in a personalized way. I even think it could help them make life decisions and come in handy with the socialization process with their friends.
Based on my existing YouTube analytics data, 24% of my audience is between the ages of 18-24, with the highest segment at 38.6% being 25+. I think this 18-24 segment will grow the most in 2022, not just for my youtube channel, but for the OpenAI community at large.
At that age, there are so many benefits of adding GPT-3 to their lifestyle. Perhaps even more so than people who may write or work in the business world professionally. I also think they are highly connected and it will be lead to very fast adoption on typically dense college campuses (even if everybody is still remote). I find younger people are more receptive to new technologies and new ways of doing things anyways.
Why haven’t we seen this segment grow already? I think the answer is quite simple - if I was still in high school, I would not have had the patience to fill out the OpenAI waitlist to try it out like in the past. Since there is no longer a waitlist, I think the easy sign up process we have now will suit a teenager’s level of patience a lot more. With this barrier out of the way, I think the GPT-3 floodgates will open to many high schoolers and college aged students who very actively need help with their homework and schooling.
Besides the risks, I was surprised to learn that today’s students are already using AI based writing tools to help them with their homework. Tools like QuillBot and Grammarly are being used actively. I wouldn’t be surprised either if Github copilot/OpenAI Codex is a must-have tool for programming students right now. I think the difference will be that GPT-3 is just so much more powerful than any other AI based writing tool on the market. This delta of improvement will lead to it having greater virality amongst college communities once they start to realize all the different ways it can help them as students.
Although technology has been entering school life more and more over the past few decades, I really do think GPT-3 represents a whole new jump of technology altogether. It’s not quite fair to compare it to something like calculators, the personal computer, mobile phones, or software tools like Google Translate. Out of the box, it has really compelling capabilities which we are still discovering every single day. A lot if its text is indistinguishable from human written text. Also, in my experience, by running the text it generates through plagiarism detection algorithms, I’ve noticed it passes with flying colours. So, even if they wanted to put a stop to it, I’m not sure if GPT-3 completions can even be detected and appropriately surfaced to teachers and school administration. At the same time, punishing students for using AI tools like GPT-3 is even harder to do nowadays, since everyone is using their personal devices in their homes and doing school remotely.
The larger question is how will schools react? GPT-3 raises important questions about the grading system, how we evaluate students’ capabilities, the future characteristics of the workforce, the role of teachers, the value of a college education, and the values of the school system itself.
Please keep in mind, GPT-3 and OpenAI products are on an exponential trajectory. This means “Math GPT-3” will get better, language based GPT-3 will get better, OpenAI Codex will help even more CS undergrads with their lab assignments, multimodal models like DALL-E may allow students to make the best art we’ve ever seen. There was even a new type of GPT-3 announced last week which can search the web and synthesize online research for you too! These increasingly powerful models represent new verticals in which GPT-3 will infiltrate our education systems. Although there may be a small group of early adopter students who benefit from it at first, I think these models may overtake the student population altogether with serious saturation and mass adoption.
I can talk in more detail about how this mass adoption of GPT-3 may play out in the school system, but really, my preference would be that the school system itself reforms in foundational ways. I have many gripes about the university acceptance process, which subjects high schools tend to prioritize, the lack of school accountability for real world student success, how courses are conceptualized and designed, the costs associated with college education, as well as the historically low levels of university acceptance rates.
Perhaps, if anything, GPT-3 represents a tool for equalizing students skills and capabilities. Even if English may not be your first language, perhaps, GPT-3 could transform your brilliant ideas into successful essays - allowing you to excel in the school system in ways you could not before. Is test performance, memorization, the ability to follow essay formats, and mathematic calculation the highest form of our students’ potential? What kind of person will society need in the future? How will students excel and achieve new heights in an AI assisted world?
I want to close by saying that I’m not sure to what extent the community or OpenAI is ready for this transformation. Both in terms of scale but also our capacity to ramp up and make fair usage policy decisions. For various other reasons, I’m just convinced this will be the fastest growing GPT-3 customer segment in 2022. By my knowledge, for a while now, I think I’m the only one in the community who has really been calling this out and saying it will be huge. There’s always the chance I could be wrong, but I do think this is a topic we should start thinking more about as an industry.
We’ve seen a lot GPT-3 products made to help marketers write better copy, but this GPT-3 in the education system category shows a lot of promise too. I guess I’m saying that the exciting part about this revolution, is that this educational transformation may actually be student led and could change education forever! I’m excited to see what all of the students come up with.
Good post! Colleges & universities are totally unprepared for this.
I think the best thing for gpt-3 for my own usage is literature search. It would be nice if I could feed it papers on a certain topic and ask it questions. Like I could feed it papers on immuno oncology and then ask it specific questions on which mechanism is important for tumor infiltration.