“`html
Illustrations by Liz Zonarich/Harvard Staff
Arts & Culture
What value does writing hold anymore?
Researchers from diverse fields discuss the significance of writing amid the emergence of generative AI technologies.
What do learners risk forfeiting if they no longer engage in writing?
Since the introduction of ChatGPT in 2022, numerous students have sought AI assistance for writing assignments, and its usage is expected to increase as students become more proficient. This transition has provoked immediate apprehensions among educators regarding academic honesty and broader implications for intellectual and cognitive growth.
The Gazette engaged with faculty from various fields, including a computer scientist, a philosopher, a neurologist, and two cognitive scientists, among others, to inquire about potential future developments. We posed the same question to GPT-4, which identifies itself as “OpenAI’s most sophisticated system designed to deliver safer and more beneficial responses.” The conversations have been condensed for brevity and clarity.

Alice Flaherty.
Stephanie Mitchell/Harvard Staff Photographer
Neurologist Alice Flaherty, M.D., Ph.D., Associate Professor of Neurology and Associate Professor of Psychiatry at Harvard Medical School
We forfeit capabilities when we delegate tasks to others or machines. Since our brains have limited capacities, they inevitably lose facts and abilities we no longer employ, creating space for new information and skills.
The London cab driver studies serve as a classic illustration of competition for resources. As they became familiar with the London road network, their posterior hippocampus physically expanded. Conversely, their anterior hippocampus diminished — and they performed worse on visual recognition tasks compared to those without intricate mental maps.
The crucial question is: When AI liberates all the neurons that are currently engaged in selecting the correct adjective or metaphor, what new capabilities will AI facilitate? We cannot foresee that. If AI becomes competent enough to perform all tasks we can, we cannot even determine if any new skills will remain for us to learn.

Leslie Valiant.
Harvard file photo
Computer scientist Leslie Valiant, T. Jefferson Coolidge Professor of Computer Science and Applied Mathematics
Education will undergo reevaluation in the forthcoming years. AI will act as the catalyst for this transformation, but not the primary rationale behind its urgency. The core reason, in my opinion, is that thus far we have not acknowledged human learning potential as a phenomenon fundamentally worthy of investigation and comprehension. Instead, education has been approached as a best-practice initiative with minimal intellectual foundations.
How does AI factor in? With AI, one can replicate any human cognitive process that can be defined and for which sufficient effort is applied. This perspective was foresightedly presented by Alan Turing as early as 1951. Recent findings indicate that disputing this idea will likely be unproductive.
Consequently, education should not be framed in terms of competition between humans and machines. The central question for educators is: What can and should education accomplish for individuals? Essays may or may not form part of this response. Furthermore, the complications of assessing student work when AI is accessible should not obscure this issue.

Susanna Siegel.
Harvard file photo
Philosopher Susanna Siegel, Edgard Pierce Professor of Philosophy
If you regard a paper as a response to an inquiry, you might assume that the sole purpose of composing the paper is to address that inquiry. If a machine can provide an answer just as effectively or even better than a student, isn’t it more effective to utilize the machine’s response?
This perspective on the situation overlooks the advantages to the learner from the exploration process. As a completed work, a paper rehearses a line of thought, sometimes serving as a pathway to an answer to a question.
However, the outcome renders invisible the…
“““html
Tangled threads of inquiry, the misleading yet captivating beginnings, the fruitful, exquisite disarray of contemplation and composition comprised of elements that once appeared pertinent but ended up being insignificant. These elements are a blend of grit and brilliance. Within this disarray lies epistemic advancement.
Ultimately, when we delegate the responsibility of discovering words to articulate thoughts, we forfeit all the forms of cognitive engagement that arise from those quests. Words and concepts are infused with associations, reminders, emotional significance — an entire landscape of links to our history and future projects, notions, and relationships.
There is also a communicative importance to the act of creating from the ground up; as you navigate someone through a thought process filled with entanglements and misdirections, you enhance their comprehension.

Tomer Ullman.
Harvard file photo
Cognitive scientist Tomer D. Ullman, assistant professor of psychology
Imagine if a wormhole appeared tomorrow and an extraterrestrial being named Bunfo emerged. At first glance, Bunfo seems unintelligent, but as conversations progress, its intelligence becomes more apparent. Its main desire appears to be answering any questions posed. You might inquire, “How do you express ‘I love you’ in Italian?” and it would reply, “Ti amo.” Ask about a pizza recipe, and it would provide one.
Before long, entire industries arise to connect devices ranging from smartphones to aircraft to Bunfo, as scholars fervently debate its thought process and desires. Soon enough, students begin asking Bunfo to compose their papers. What do these students forfeit? They miss out on whatever writing essays is intended to impart.
Sure, if your goal is to order pizza in Rome, it’s entirely logical to present a menu to Bunfo and inquire, “What does this mean in English?” However, if you’re enrolled in a course to learn Italian for personal growth and you only rely on Bunfo for answers, then what purpose does your presence serve?

Talia Konkle.
Photo courtesy of Talia Konkle
Visual cognitive computational neuroscientist Talia Konkle, professor of psychology
Some pertinent insights from cognitive science pertain to active versus passive learning. Active learning — creating knowledge independently — enhances retention. For instance, the widely recognized “testing effect” is a potent phenomenon in educational research; the act of assessment, of attempting to formulate answers, fosters retention.
A straightforward illustration is vocabulary acquisition. If you read a word alongside its definition, your retention won’t be as strong compared to if you view the word alone, pause, attempt to guess, and then consult the flashcard for the definition afterward.
When we enlist AI for active learning, I see significant advantages. However, if we employ it to bypass our critical thinking regarding the skills we want to internalize, that could be detrimental.
I can anticipate when I’m engaging in work that leads to new insights, where it’s essential that these insights embed in my cognition. This awareness impacts how I utilize AI in my learning approach. I also recognize circumstances when the output I need is less tied to my educational aspirations, where I’m glad to use AI’s synthesizing and summarizing capabilities for efficiency.
This is where the essence lies in aligning students’ learning objectives while composing an essay. We, as educators, can convey our expectations of their takeaways from the writing experience and define what we consider successful. However, just like with any task, students can engage more or less actively with that undertaking. AI may simplify the process or entice passivity among learners, yet I believe it can also enhance our efficiency as proactive learners.

Joshua Greene.
Photo by Dylan Goodman
Experimental psychologist, neuroscientist and philosopher Joshua Greene, professor of psychology
When students allow AI to undertake their cognitive tasks, they forfeit the opportunity to enhance their critical thinking abilities. What these students sacrifice now parallels what previous learners lost by hiring help from friends, family, or paid assistance. What differentiates the current scenario is the expectation that this alternative intelligence will remain accessible indefinitely.
I contend that it remains vital for individuals to cultivate strong critical thinking skills; however, as machines become more advanced, many will question: Why put in the effort? My response: Even if the demand for human cognitive output significantly lessens, we still need wise individuals at the helm.
This necessitates that some — ideally all — humans retain the intellectual capacities needed to conceptualize and actualize a favorable future. In the coming years, the demand for lucid thinking may evolve into a more democratic necessity rather than a technocratic one.

Mary Lewis.
Photo by Dylan Goodman
Historian Mary Lewis, Robert Walton Goelet Professor of French History
The aim of writing is not solely the outcome but also the journey. Students who solely rely on AI for crafting their essays forfeit the essential thinking process.
“`
And that is, ultimately, what we aim to achieve: generate new insights. If the objective is to engage in thoughtful reflection or create novel knowledge, then only a human can truly accomplish that.
Inherently, AI reprocesses existing material. How many of us have experienced instances where, while messaging someone, AI attempts to anticipate our next statement but fails? Why does this occur? Because it relies on what it predicts is the most probable continuation of our conversation. Hence, in a document, if AI employs predictive text, it will merely recycle pre-existing concepts rather than inventing original ones.
The greater the volume of information input into AI, the more data it has at its disposal. As this unfolds, it may appear as though it is reflecting, but in reality, it operates within a vast, closed loop. Therefore, I reiterate my initial assertion: Writing catalyzes thinking; it is not merely a finished product, but a practice that ignites contemplation.

GPT-4, “OpenAI’s most sophisticated platform that delivers safer and more beneficial replies”
When learners permit AI to compose their essays, they forfeit the critical education gained from interacting with the writing process. Writing fosters analytical thinking, deepens comprehension of the subject, and enhances communicative abilities.
By circumventing this experience, learners miss the chance to clarify their own thoughts, understand how to construct arguments, and cultivate their distinctive voice. They also jeopardize a genuine understanding of the material, given that significant learning often occurs during the challenge of organizing and articulating ideas clearly.
Moreover, utilizing AI to generate originally intended work can raise ethical issues and impede students from receiving valuable feedback that could aid their development. Ultimately, complete dependence on AI can undermine both academic growth and personal self-assurance.