What Does It Mean to Be Written by a Human?

ChatGPT became available to the public on November 30, 2022. Within just five days, there were one million users, and within two months, 100 million. At the time, I was in my second year of my EMBA at MIT, surrounded by highly motivated and technologically savvy classmates. In 2022, it was considered unethical to use artificial intelligence (AI) to complete assignments. Still, there was tremendous enthusiasm about how to use AI in many industries, including healthcare, manufacturing, social media, customer service, software, and law. At the time, I also had a college student and a high school student, and educators at all levels were banning the use of any AI, as using AI was considered cheating.

Four years later, our world has been forever changed by AI. We use AI note-taking apps to decrease the burden of clinical notes, and many companies use AI agents for customer service calls for everything from healthcare to hospitality. There is concern over the environmental damage and job losses caused by data centers. Bill Gates recently wrote an essay on his website, “The turbulent AI era is here. The choices we make now are critical,” sounding the alarm about the dangers of AI given the extreme pace of innovation and “lack of a plan to ease AI entry” (Gates Notes. https://tinyurl.com/5t4kh294).

So, what does it mean to be written by a human?

The Authors Guild initiated the “Human Authored” project, allowing authors to certify that their work was written by a human and not generated by AI, if only AI applications that check spelling and grammar or help with research were used. Currently, most agree that it is unethical to type into ChatGPT, “Write a 4,000-word essay on Renaissance art,” and turn it in as one’s own, or to copy a manuscript into ChatGPT and ask it to review the manuscript, but there are many shades of gray between having AI do all the work and using AI to learn, educate, and improve the quality of the written document.

Currently, college professors acknowledge that students will use AI. The goal of the class is for the student to learn the material, so they have changed their grading structure. Homework is graded for completion, and tests taken in person, where AI cannot easily be used, count for most of the grade. Oral exams have also been initiated, so students have to present in person, in their own words, knowledge about a topic.

I have been a slow adopter of AI, but over the past few years, after many discussions with my friends in law and biotech and attending AI conferences, I have become an adopter, as there are areas where I find AI to be very helpful. In the past, I might have had my husband or a friend review an email for tone, but now I will have ChatGPT or Copilot do this. Does this mean that I did not write the email?

A friend of mine in biotech uses ChatGPT for presentations and as a strategic coach. This is highly supported by her company. She feeds ChatGPT all of the needed data and then checks the output. This saves an incredible amount of time, shifting the time previously spent on slide creation to the analysis of different strategies, something where human input has maximum utility. Felipe Csaszar wrote about how new tools can improve human judgment by tirelessly generating, evaluating, and synthesizing insights in his recent article in Harvard Business Review, “AI Is Revolutionizing Strategic Decision-Making” (HBR. https://tinyurl.com/2cd9yt6y). He makes the argument that AI use can free up human bandwidth so that humans can use their time for critical thinking.

Currently, there are more questions than answers regarding the use of AI, but AI is not going away. Asking the difficult questions and designing governance structures now is critical to “ease AI entry” and attempt to make the good of AI outweigh the bad. Human authorship may ultimately be less about who typed every word and more about who generated the ideas, exercised judgment, verified accuracy, accepted responsibility, and had something meaningful to say.

What Does It Mean to Be Written by a Human?

AI-edited editorial. My draft was cut from 760 to 500 words with any edits that AI wanted to make.

ChatGPT became publicly available on November 30, 2022. Five days later, it had one million users; within two months, 100 million. At the time, I was in the second year of my EMBA at MIT, surrounded by highly motivated, technologically savvy classmates. There was tremendous enthusiasm about AI’s potential in healthcare, manufacturing, law, customer service, and countless other industries. Yet using AI to complete an academic assignment was considered cheating.

Four years later, the distinction is far less clear.

AI now helps clinicians document patient encounters, companies interact with customers, scientists analyze data, and professionals create presentations and written materials. At the same time, legitimate concerns exist about job displacement, environmental costs, misinformation, and what happens when technological innovation outpaces our ability to establish appropriate guardrails.

This raises a deceptively simple question: What does it mean for something to be written by a human?

The Authors Guild’s Human Authored initiative allows authors to certify that the text of a book was written by a person rather than generated by AI, while permitting limited uses such as spelling, grammar, and research assistance. That definition seems straightforward—until we consider how people actually work.

Few would argue that asking ChatGPT to “write a 4,000-word essay on Renaissance art” and submitting the result constitutes human authorship. But between writing every word unaided and asking AI to create the entire product lies an enormous gray zone.

I have been a relatively slow adopter of AI. In the past, I might have asked my husband or a colleague to review an important email for tone. Today, I may ask ChatGPT or Copilot. Did I still write the email?

A friend in biotechnology uses AI to help develop presentations and, as a strategic coach, a practice encouraged by her company. She provides the data, interrogates the output, corrects it, and ultimately decides what conclusions to present. AI reduces the hours spent constructing slides and allows her to devote more time to evaluating strategies and making decisions. Which part of that work matters most?

The same questions increasingly apply to medicine, research, education, and scientific writing.

Perhaps we need to reconsider our definition of authorship. Human authorship may ultimately be less about who typed every word and more about who generated the ideas, exercised judgment, verified accuracy, accepted responsibility, and had something meaningful to say.

AI can generate, organize, critique, and synthesize at extraordinary speed. Used thoughtfully, it may allow us to spend less time on tasks machines perform well and more time on those requiring curiosity, empathy, creativity, ethical judgment, and wisdom.

There are still far more questions than answers. But AI is already here. Our responsibility is not simply to decide whether to use it, but to determine where human judgment must remain indispensable.

Perhaps that is what being “written by a human” should come to mean.

What Does It Mean to Be Written by a Human?

Independently written by AI with the prompt “Write a 500-word editorial on what it means to be written by a human. Do not use my draft for content”

For most of history, the phrase “written by a human” would have been absurd. Who else would have written it? A novel, scientific paper, newspaper column, letter, or hastily composed email necessarily originated in a human mind. Technology might have changed how words reached the page—from quill to typewriter to keyboard—but authorship itself was never in question.

Artificial intelligence has changed that assumption.

Today, an algorithm can produce a polished essay in seconds. It can summarize complex ideas, adjust tone, suggest metaphors, correct grammar, and imitate nearly any style. Increasingly, it may be impossible to determine from the words alone whether a person or a machine produced them. This has led to understandable concern about authenticity, originality, and even the future of writing.

But perhaps we are asking the wrong question.

The important distinction may not be whether artificial intelligence participated in producing the words. After all, writers have always used tools. We consult dictionaries, editors, colleagues, search engines, and reference materials. Few would argue that spellcheck diminished human authorship. The more meaningful question is: Where did the thinking come from?

Human writing begins before the first sentence appears. It begins with experience.

We write from what we have seen, learned, loved, feared, regretted, and hoped for. A physician writing about medicine brings memories of patients whose lives changed how she practices. A scientist brings years of failed experiments alongside the one that finally worked. A parent writing about childhood carries memories unavailable to anyone else. Even
when those experiences never explicitly appear on the page, they influence which questions we ask, what we notice, and what we believe matters.

That is what makes human authorship more than the mechanical production of sentences.

Artificial intelligence complicates this because it is an extraordinarily powerful collaborator. It can challenge our assumptions, help organize ideas, expose weaknesses in an argument, and sometimes find language for something we have struggled to express. Used thoughtfully, it may make us better writers.

But there is a danger in allowing efficiency to replace thought.

Writing has never simply been a method of communicating what we know. Writing is one of the ways we discover what we think. The frustrating search for the right word, the paragraph rewritten five times, and the argument that changes as we struggle to articulate it are not inefficiencies to eliminate. They are part of thinking itself.

Perhaps, then, “written by a human” should not mean untouched by artificial intelligence. Such a standard will become increasingly difficult—and perhaps increasingly meaningless—to maintain.

Instead, it should mean something more demanding.

It should mean that a human being remained responsible for the ideas, judgment, perspective, and truth of what appears on the page. That the writer questioned rather than merely accepted. That the final words represent something the author genuinely believes and is willing to stand behind.

In an age when words have become remarkably easy to generate, the value of writing may increasingly lie not in who typed the sentences, but in who did the thinking.

And that responsibility remains profoundly human.          

  —Robin