John Henry vs AI

I have been having a discussion on a Mastodon instance about the knowledge base supporting AI apps. It was initiated by someone who proposed that if AI is based on the content of too many content creators developing content using AI wouldn’t the improvement of the knowledge base bog down rather than become more helpful. This made some sense to me.

Imagine the following comparison. I am a retired academic and in retirement, I have transitioned from doing research to writing based on my reading of research. This is where the John Henry reference came from. Instead of John Henry the “steel driving man” against the steam engine, it is Mark the blog-writing man against ChatGPT. I read stuff and then write stuff. ChatGPT scans stuff and then with prompts writes stuff. We both can now only build from the knowledge base that exists and then contribute back to that knowledge base. Neither ChatGPT nor I can contribute insights back to that knowledge base. We both can summarize and interpret, but not hypothesize and test.

Granted ChatGPT has the potential to access more content than I can possibly read. I think I can speculate and propose in ways that ChatGPT can not, but I admit I cannot test my speculation.

Without new inputs, we may get to a point similar to that of those Republicans who suggest that Fauci lies. I would argue that a scientist being wrong as viewed in hindsight is probably not proof that a scientist has lied. It is that a scientist may have been relying on the existing data and made an adjustment when new information revealed that an old idea was wrong. Likewise, Chat GPT needs new and independent inputs to the knowledge base to advance.

I have been exploring a related idea by using ChatGPT to comment on topics I know to be based on research reaching conflicting decisions. An issue within my own domain of study that fits this scenario is the question of whether college students should take class notes using laptops or paper and pencil. Perhaps this is not an issue many see as earth shattering, but the point is here is an issue that is unresolved yet practitioners are relying on positions taken to make decisions. I land on the laptop side because of the long-term benefits I see in collecting notes in a way that allows the learner to make use of these notes in more powerful ways over time and in the future.

The more common position at present is that taking notes by hand is more productive. Some studies do indicate this is the case, but I keep asking for a rationale that would explain the advantage. The best existing explanation I know of goes something like this. Most students can enter text on a keyboard at a significantly faster rate than they can write by hand. On the surface, this might seem to be an advantage for laptops. However, the argument speculates that because of limitations pen and paper users think more carefully about what they will record and benefit from this more active cognitive behavior. Transcription or close to it involves less thinking. Ok, perhaps, but this position ignores that keyboarding is under the control of the laptop user and that they could use the more efficient input to think more and record the same amount as the paper users. Does this seem reasonable?

So, I presented this “issue” to ChatGPT to see what it would say. Here is what I got.

Is it better to take notes on a laptop of by hand on paper?

Both taking notes on a laptop and by hand on paper have their advantages and disadvantages, and the better option depends on personal preferences and the situation at hand. Here are some factors to consider:

– Faster typing speed: If you are a fast typist, you may be able to take notes more quickly on a laptop than by hand, which can be helpful if the speaker is speaking quickly or if you need to take a lot of notes in a short amount of time.

– Searchable notes: With digital notes, you can use the search function to find specific information quickly.

– Easy editing: You can easily revise, reorganize, and format your notes without having to cross out or erase anything.

– Convenience: You can have all your notes in one place and access them from anywhere if you save them to the cloud or a digital note-taking app.

Advantages of taking notes by hand:

– Better retention: Some research suggests that taking notes by hand may lead to better retention and understanding of the material compared to typing. This is because handwriting engages different parts of the brain and can help with processing  and remembering information

– Less distracting: When you take notes by hand, you are less likely to get distracted by notifications or other apps on your device.

– More flexibility: You can draw diagrams, arrows, or other symbols easily by hand, which can be helpful for visual learners or when taking notes in math or science classes.

Ultimately, the better option for you depends on your personal preferences and the situation at hand. If you are more comfortable with a laptop and need to take a lot of notes quickly, then typing may be the better choice. If you prefer handwriting and want to retain the information better, then taking notes by hand may be the way to go.

Can you provide citations I should read on this topic?

Mueller, P. A., & Oppenheimer, D. M. (2014). The Pen Is Mightier Than the Keyboard: Advantages of Longhand Over Laptop Note Taking. Psychological Science, 25(6), 1159–1168. https://doi.org/10.1177/0956797614524581

Carter, M. J., & Van Matre, N. H. (2016). Note-taking and Handouts in the Digital Age. College Teaching, 64(2), 67–77. https://doi.org/10.1080/87567555.2015.1124408

Kiewra, K. J., Benton, S. L., Kim, S., Risch, N., & Christensen, M. (2019). Investigating the Benefits and Drawbacks of Digital Versus Traditional Lecture Capture. Journal of Educational Psychology, 111(5), 725–741. https://doi.org/10.1037/edu0000315

Kontra, C., Lyons, M. J., Fischer, S. M., & Beilock, S. L. (2015). Physical Experience Enhances Science Learning. Psychological Science, 26(6), 737–749. https://doi.org/10.1177/0956797615572659

I hope you find these resources helpful!

Mark the blog-writing man cannot compete with this. I have read several, but not all of these references. ChatGPT hedges, because students either use one system or the other. The specific comment about retention claims it is best with hand written notes. The “Pen is mightier than the keyboard” study is frequently cited to support requiring students to use pen and notebook instead of a laptop. I described this study in a previous post. I believe I have an advantage in being able to add analysis to this issue.

The issue is whether the knowledge base ChatGPT or me uses will improve as new evidence is translated into content. This is a complex challenge that depends on continued incentives for researchers and those who work to translate and share what these researchers contribute. Any change in the incentives supporting these individuals could reduce the improvement of the knowledge base that guides human behavior. ChatGPT and other AI systems ultimately rely on a knowledge base humans built. Making summarization and communication easier or less costly does not change the need for the constant upgrading of this knowledge base.

Loading