Dr. Ayse Ozturk

Note: To keep that precious human imperfection, this essay got zero help from AI or any other assistance tools (except for the news links/excerpts). Apologies in advance for any typos or errors, but I hope the 100% authenticity, which is increasingly harder to find in the age of AI, makes up for that.

---

I’m an educator, a college professor. Since childhood I kept my curiosity on many things, and for the last two years it has been focused on AI. As I got to learn more on AI, I got more excited, cautious, hopeful and pessimistic all at the same time. My main responsibilities include teaching and research, which are significantly affected by AI. There are revolutions, major shifts happening in teaching and research and on the job market due to AI progress. It is at a dizzying pace, hard to catch up, which adds to the complications of the challenges it poses. But here are the three existential issues that I believe the educators should pay attention to find an answer or a roadmap in the face of AI.

1. Everyone now has a personal tutor, which is great but wait? What about the professors and their role in education?

Watch the video below on Google’s Project Astra to see where education is going. Everyone now has a personal tutor. Isn’t this a good thing? Yes, mostly. But these tools come with many downsides such as hallucinations, making us lazier, curbing the learning process by shortcutting to the answers. Well, human instructors may also come with similar downsides. So, overall having personal digital tutors must be a good thing. Then, what is the new path the education is heading toward? Is it each student served with a digital personal tutor? And a cental entity to coordinate the progress of students? Where in this picture do the roles of the current actors fit in? Will there be no more human professors? Or will there be less of them? Will the professors act only as mediators between digital tutors and students? Currently, it can be a co-creation process between a human professor and a digital tutor to train/educate the student. But what happens in the long term? If Artificial General Intelligence (AGI) or Artificial Superintelligence (ASI) is achieved at one point, what would justify the role of human professors? Some argue that we have already reached the AGI stage, which I believe is partially true (because it is a vague concept anyways), then human professors can only serve for human connection and additional enhancement. But when ASI is achieved, meaning AI surpasses human intelligence in all aspects, then the teaching job, along with many other knowledge industries, are going to look very much different than how they look today. We’ll wait and see. But in the meantime, let’s ponder over what would be our meaningful role in the age of AI that provides personal digital tutors for everyone.

2. The job replacement/displacement for entry-level positions

We, at higher education institutes, teach and train students to get them ready for the workplace. We focus on multiple facets of learning, i.e., theory, application, critical thinking, analytical thinking, problem solving, communication, writing skills, etc. However, the recent news on the gloom of entry-level positions should make us question our traditional approaches of teaching. What should we be doing now if employers can just use AI to get many entry-level tasks done? Some companies publicly announced that they would stop hiring unless AI can’t do the job. For example, Duolingo has shifted to an “AI-first” approach; Shopify CEO has instructed teams to demonstrate that a task cannot be done by AI before requesting additional hires. We should go deeper or think out of the box to identify the problem and come up with creative new approaches. It includes upskilling the educators so that they are able to pass those new skills on to students for the emerging needs of the job market. See the AI shift on the tech job market below.

• Nearly 25% of U.S. tech job postings this year are looking for candidates with AI skills, according to job listings data. Companies across various industries adapt their hiring strategies to integrate the technology.

Source: WSJ, How the AI Talent Race Is Reshaping the Tech Job Market, March, 2025.

Here are some articles about the entry-level job displacement problem:

• No hire, no fire: The worst job market for grads in years: The unemployment rate for recent college graduates has climbed by 1.6 percentage points since mid-2023, and entry-level hiring is down 23% compared to March 2020. 

• Dario Amodei, CEO of Anthropic, warns that AI could wipe out all entry-level white-collar jobs and spike unemployment to 10-20% in the next 1-5 years. The CEO suggests that to mitigate the worst scenarios, the government and AI companies should increase public awareness, help workers understand how AI can augment their tasks, and debate policy solutions such as job retraining programs and taxes on AI companies to redistribute wealth.

• Early evidence of AI-led entry-level-job apocalypse: For Some Recent Graduates, the A.I. Job Apocalypse May Already Be Here: Unemployment for recent college graduates has jumped to 5.8% as companies replace entry-level workers with AI. AI is automating white-collar jobs, raising concerns about underinvestment in training and mentorship for young workers. Some graduates are pursuing riskier career paths to stay ahead of AI’s impact on traditional jobs.

• A.I. Is Coming for Entry-Level Jobs: The unemployment rate for college grads has risen 30 percent since September 2022, compared with about 18 percent for all workers. AI poses a real threat to entry-level jobs, as advanced tools automate tasks once done by junior workers. This coincides with rising unemployment among college grads and declining job confidence, especially among Gen Z. To address this, employers must redesign entry-level roles to provide higher-level, value-adding tasks, while education providers must align curricula with emerging skills needs.

• Which jobs will artificial intelligence augment and which will it automate? The Anthropic Economic Index, based on anonymized data from Claude.ai, tracks AI’s impact on the job market. Early findings show that AI is currently impacting highly digitized jobs the most—especially in fields like coding—while roles requiring complex human interaction, like nursing, elder care, and teaching, are least affected. Related: Microsoft has laid off approximately 6,000 employees, with a significant portion being software engineers, as the company shifts its focus toward expanding AI infrastructure and moves away from traditional development roles.

• Nearly 20 million jobs are on the chopping block to be replaced by AI, SHRM research shows.

• Will AI replace humans? AI can augment human capabilities rather than replace jobs, if we design it to be a “bicycle for the mind” that enhances our abilities, rather than just automating tasks. MIT economics professor Sendhil Mullainathan says it is in humans’ power to put AI on a path to help us rather than replace us.

• My blog post: Is AI a skill-leveler or -divider? The answer is it depends on the nature of the task.

3. Research: Ethics, Credits and Breakthroughs

I watched this video today: One-Day Research Paper? AI Tools Turn Data Into Gold. The video shows how to use a few AI tools to write an academic paper from scratch and mentors even till the post-publication steps. I appreciate Andy Stapleton for bringing these use cases to our attention. It is very helpful to know these but it also brings up so many questions and calls for action.

My questions are a long list. However, the main topic is that the current state of research does not go well with what AI offers. The academic research is supposed to be innovative, original, contributing to the creation of new knowledge and is eventually an intellectual property of the researcher. AI is more or less covering the first three items; it can generate innovative, original (debatable but maybe), and contributing work as much as an average researcher can. This is maybe not at the Nobel prize winner level, breakthorugh research (yet) but the video shows it can do a great job that would pass the first round of many journal review processes. But whose intellectual property is it? Most journals ask for a disclosure of AI use; but how can you disclose in this case? If writing a paper as in the video is not ok, then who would enforce that, especially when there are no control mechanisms to do that?

And here goes my long list of questions on the topic:

• How to check if the author uses AI if the AI-detection tools do not work with total accuracy (which currently seems to be the case)?
• How to control it if a reviewer uploads a working paper on AI whereas they are not suppsoed to do?
• Even if people publish today using AI work and these work slip by the editors and reviewers unnoticed today, there is a chance that AI-detection tools will improve and can identify these work retrospectively. Will many papers be retracted by then?
• The lack of control mechanisms may be pushing many people to be more cynical of the reliability and the prestige of the academic research process.
• We can’t use AI to create an entire paper and we can’t assume that we should not get any AI help while writing a paper (e.g., using Grammarly spell check, especially for non-native speakers). Then, what is the right level of AI-help a researcher can get? Who decides that? If individuals decide that and it all boils down to personal ethics, then the least ethical ones will be the ones with an advantage and the most ethical ones will be at a disadvantage.
• If we can’t trust anything we read and approach it with suspicion that it is AI-written, then the importance of oral communication will inevitably rise. Academics will need to prove their expertise at conferences, and maybe many more conferences allowing for deeper discussion formats should be developed (e.g., the researcher presents for 30 minutes and discusses for another 30 minutes to prove the domain expertise instead of the typical 15-20 minutes, at least in my field of marketing/business). Doctoral students will be pushed to prove their work orally rather than on a written context. The same way, instructors will need to evaluate less of the written work and more of the oral work when grading the students.

I bet if we ask an AI, it will come up with many further questions beyond the ones above. But no, I’m resisting the urge…

---