Graduate studies in any field have long had a requirement for students to complete a final project to demonstrate their mastery of the subjects they’ve studied. Traditionally, this last hurdle came in the form of a master’s thesis: a hundred pages of logic and reasoning straight from your soul for a committee of professors to gaze into and assess your progress.
In recent years, many fields have either allowed or required the thesis to be substituted for a capstone project instead.
In either case, the ultimate goal is to put students through an obstacle course of research, original thinking, and expression of those thoughts in a way that demonstrates they have developed a command of the concepts and practices that make up their field.
To that end, it’s an intensively individual process. While advisors may provide guidance, and research assistants may turn up material, all the thinking has to happen in the mind of the student.
Artificial intelligence may offer another option.
You’re Joining a Field That May Change the World… And You’ll Be Among the First to Experience Those Changes
In other kinds of degree programs today, the very suggestion that AI might help develop your culminating project might raise hackles and start some fiery ethical debates.
In an AI master’s program, though, it’s just part of the package. After all, you’re supposed to be breaking new ground in applied uses of AI. After all, using it to augment your own imagination to deliver your best version of your final project is right in line with the promise of AI.
Still, it’s a tricky business. Even in a field stacked with machine learning, it’s the human mind that ultimately needs to learn the most. Your culminating project is an expression of the sum of your education, your skills, and your original ideas in the field. It can’t simply be an ML-generated stack of pages spit out by an algorithm.
And that’s not what you want, either. A master’s degree in artificial intelligence is a substantial investment in your education. It’s a first step into a field awash in theory and concepts that even the brightest minds wrestle with. If you’re not maximizing your studies to shape your own mind, you’re not going to get far as an AI engineer.
So your big challenge will be to find a way to use AI tools to support your thesis or capstone project in ways that build your skills rather than blunt them.
First Step: Understanding the Ethical Implications of Using AI in Your Master’s Thesis or Capstone Project
The very first thing you’ll have to do if you are interested in using AI to support your thesis or capstone efforts is to figure out if you should.
And that might be the most difficult part of the process.
Of course, the ethical objectives are really just the flip side of the issue we just noted about maximizing your own learning through the thesis project. Solve one problem, and you’ve solved them both. That’s a big part of the reason why there are ethical guidelines in academic writing in the first place!
How the Ethics of AI and Education Are in Alignment
In general, the historic perspective on using computers and applications in theses and other academic work seems to have been to permit them so long as they aren’t substituting for mastery of the core skills being assessed.
After all, no one bats an eye today if you run your manuscript through Grammarly for suggestions and corrections before you submit it. And you would have to work very hard to find a word processor to type on these days that doesn’t automatically correct your spelling as you type. Going further, many have incorporated autocomplete… a statistical system for predicting the likeliest next word in your sentence and inserting it automatically.
Not a few people have compared AI text generators to hyper-advanced autocomplete systems, and for good reason.
So one question to ask is when does the statistical calculation rise to the level of artificial intelligence?
You wouldn’t be interested in the field in the first place if you weren’t interested in the answer to that question. And it won’t surprise you to learn there’s little consensus on the subject just yet.
AI Tools Are Putting University Ethics Committees in a Tough Spot
You can see university plagiarism policies already trying to dance on the head of this pin already. While clarifying that plagiarism is the act of representing another’s work as your own without offering appropriate credit, regardless of how that work was obtained, many then go on to justify grammar and spelling checkers explicitly because it is your own writing being scanned rather than the output of an AI.
But at what point does it stop being yours and start belonging to the algorithm? The instant you hit “Accept All” and allow a grammar checker to reword your paper, those words stopped being, in the strictest sense, your own. And if university ethics experts are fine with that, then where are they drawing the line on ChatGPT?
These policies also get tripped up quickly when it comes to considering the sourcing of the generative transformers. One school offers the objection that “…even when AI generative works don’t directly plagiarize, they often reuse ideas and steal core concepts from copyrighted works without crediting the original creators. This is still considered plagiarism.”
Yet that sounds very much like the process of synthesis that you will be required to undertake as a grad student writing your thesis paper – even by hand! It’s the intended outcome of your studies.
Expect to see more changes in these policies soon… and pay attention to how they may affect your plans.
Original Ideas and Effort Are the Key to Using AI to Maximize Returns on Your Final Project
A key promise of the thesis experience has always been that it represents original thinking on the topic.
So, maybe the key question of using AI for thesis or capstone projects is whether that work still represents originality. Or is the output of tools trained on millions of pages of other people’s work inherently duplicative, if not outright plagiarism?
Part of demonstrating a mastery of artificial intelligence at the level that qualifies you for a graduate degree should include an in-depth understanding of how current tools in the field work.
Every thesis comes with a declaration statement that attests to the fact that it represents your own unique work done exclusively in pursuit of your current degree program.
That puts the idea of using AI without qualification right out the window—it absolutely represents work that was done on the topic and would need to be declared.
Of course, the major questions of using AI in your thesis or capstone project should be addressed well before you get to the point of writing up your declarations page. It’s an absolute must to discuss the what and the how with your supervisor and committee.
Understanding the Differences Between a Capstone Project and Thesis Paper
Before we get to step two, it’s a good time to talk about final project requirements in AI master’s degrees.
Artificial intelligence graduate programs span the gamut of theoretical and applied studies in the field.
Like other graduate programs, many of those on the applied side of the field have started to move toward allowing or requiring capstone projects in place of a thesis. In other cases, students are offered a choice between the two paths.
What’s the difference between these two types of final project? Understanding what they share and what separates them will help you choose AI that can best support your decision.
Features That AI Master’s Thesis and Capstone Projects Have in Common
What they have in common is the intent that they serve as a synthesis of your studies. And, as noted, they should also offer a glimpse into your own thinking on the subjects. Unlike tests or other types of academic papers, it’s not enough to just regurgitate what you’ve read. A final project of any sort is supposed to tie together what you have learned in an original expression of thought that contributes in some measurable way to the base of knowledge in artificial intelligence.
The two types of projects also share other attributes:
- Fully incorporates the current state of research on the topic available
- May include some level of original research or investigation, though not to the level of a doctoral dissertation
- Must be decided upon in coordination with, and approved by, your advisor or instructors
- Will be evaluated by a committee in an agreed-upon presentation format
In both cases, your instructors will have significant input throughout the process. While it’s ultimately your work, it’s also the case that your performance reflects on them. So they have every incentive to ensure you are picking a reasonable topic and exercising the full range of your abilities to investigate it.
Particularly in AI research, there is less and less distinction between the capstone and thesis models. Capstone projects still involve substantial writing; thesis papers often take on practical questions and involve code development.
Capstone Projects for Artificial Intelligence Master’s Programs
Capstone projects are more practical in nature than a thesis paper. They may be aimed at:
- Producing a specific, working implementation of your theory
- Evaluating existing tools, processes, or theories in AI through actual experimentation
- Looking at and solving specific, real-world problems in a practical fashion
One critical difference between capstone and thesis projects is that, at some universities, a capstone project may be shared between multiple students as their culminating program effort. This allows students to take on larger and more ambitious studies than they would be able to individually. It also demonstrates an ability to collaborate in a way that an individual project cannot.
Although it’s called a capstone project and not a capstone paper, you can still expect to do plenty of writing as a part of the process. Although the paper that you generate will be shorter than a full thesis, the writing will still have to be tightly constructed and offer convincing support for your academic objectives.
One final difference between capstone projects and thesis papers is that a capstone is almost always graded as a presentation.
In some cases, your capstone presentation may be expected to take the form of a video or other kind of multimedia format. That can require a different sort of design and production effort than a formal paper.
There is overall more flexibility in format and presentation allowed than with thesis papers.
Thesis Papers for Artificial Intelligence Master’s Programs
Thesis papers are traditionally anywhere from around 40 to 100 pages in length. There are often variations in the typical length depending on the field of study; for thesis papers in artificial intelligence, thesis text tends to run to the shorter side, from 30 to 70 pages.
These papers must conform to a more rigid format than capstone presentations. Each university typically sets out rigorous guidelines and standards. The structure of a thesis paper is often closely defined, setting out a sort of stair-step approach to describing the topic, analyzing the research, and presenting your findings.
Thesis topics aren’t required to be theoretical, but they are often launched in a general spirit of inquiry rather than aimed at a concrete solution. The focus requires picking a reasonably narrow topic of investigation; narrower, for example, than what you might tackle in a multi-person capstone project.
On the other hand, you can be sure that a thesis expresses your individual thoughts and reasoning on the subject. Although you will certainly kick around your ideas with advisors and peers, at the end of the day it’s your name on the cover page.
Second Step: Determining the Aspects of Your Project Where You Can Enlist the Help of AI
The good news is that this list is expanding all the time. New tools and possibilities are being invented as the technology evolves.
It’s hardly a foregone conclusion that AI will find acceptance in the wider academic community for thesis writing. But based on the number of companies that are spinning up to offer AI tools for that purpose, it’s clear that a lot of people are betting that it will.
You will find, or can build, tools that help you build your thesis from start to finish.
Generating Thesis or Capstone Project Topics
For many students, one of the hardest parts of a final project is getting started. AI offers new options for brainstorming research ideas and paths of investigation.
Writing Your Project Proposal
AI tools also help hone the all-important thesis proposal to your committee. With the rigor of logic and the creativity of Large Language Models, AI can help you quickly distill your methodology, explain the significance, and set the context for the project. Separately, they can assist you in any required literature review before you begin.
Developing a Project or Paper Outline
AI is a superb outliner. With rigorous structural standards and exacting completeness, it can help you organize the overall presentation and structure of your paper for maximum clarity and impact.
Conducting Research
Research is one of the areas where AI holds the most promise in thesis writing but currently holds the greatest risks. While LLMs are superb at unearthing hidden gems of relevant information that might not otherwise come to your attention in large volumes of writing, they are also currently prone to hallucinations: inventing imaginary but real-sounding facts and papers. But with very careful oversight, AI can be a boon to your literature review, citation production, and data analysis.
Computation and Analysis
The meat and potatoes of your final project work will be to prove or disprove your hypothesis.
For many AI degree final projects, particularly capstone projects, a lot of computational analysis will go into the results. This may be the place where the use of AI may be both most conventional and most useful. In fact, it is probably the point of the entire project.
You’ll probably be limited in the off-the-shelf tools available, though. Since you are engaging your topic in an original way, there really shouldn’t be existing tools that can handle it. You may start with existing libraries and toolsets, but you’ll certainly be customizing them or using them in novel ways to look for the results you need.
The actual ways you put AI to use in this role are limited only by the imagination of your proposal. Some recent work has included:
- Training Child AI Models from Existing LLM chatbots
- Autonomous AI-Training for Robot-Assisted Sorting for Efficient Recycling
- Creating a Novel Evolutionary Artificial Life Environment
Writing and Revising Your Paper
We’ll look further into the how and why and how much of using AI as a partner in writing in the next section. But it’s a no-brainer to use AI tools for the process of formatting and revising to current academic paper standards. Many students who are non-native speakers use common translation tools. Many departments even provide recommendations for citation generators!
Third Step: Getting the Most from AI to Support Each Phase of Your Thesis or Capstone Project
While there are a lot of possibilities for ways you can use AI to support your final master’s project in almost every phase, the hard part is deciding how.
This is where you make the tough choices between just using AI to make the project easier versus ways to make it better. And those choices may be entirely different for different projects.
We surpass the AI by standing on its shoulders.
Most of these processes can be done with conventional GPT LLMs, and don’t require special tools. On the other hand, new AI tools developed specifically for academic needs are emerging all the time. So you might find better options than just plain-old ChatGPT.
Generating Ideas
Brainstorming final project ideas is a timeless preoccupation for grad students. But while it’s still valuable to sit down and kick around ideas with your advisor and other students, now AI offers another option.
Feeding prompts into LLM chat bots, up to and including asking them to review existing research in the field, can allow you to kick ideas around with yourself. AI can take what you put in and turn the concepts around to uncover new inspiration and thinking in your own mind.
And if you already have ideas, but lack the precision and rigor needed to turn them into an acceptable thesis statement, AI can bridge that gap. Grammarly, for instance, which many students already use to check papers, even has an AI Thesis Statement Generator you can use to spark specific topic ideas out of a general prompt for certain audiences.
Outlining
LLMs are also excellent at breaking out ideas into outlines that you can then use as part of your proposal and a jumping off point for your research and writing. By feeding in your title and key points of the project, you can get a point-by-point tick-tock of topics that will have to be covered in the final paper.
Many GPT tools have the benefit of having been trained on thousands of existing research papers, so you’ll often get back very conventionally accepted outlines that expand on your points in ways that are typical. But these may not always reflect your thinking or goals. That can be both positive and negative, though!
By diving through existing papers, the AI may surface points to cover in your outline that you hadn’t considered—but should!
Outline generation is most valuable for getting your topic broken down into bullet points and neatly summarized, section by section. While it’s possible to continue prompting and working through your outline with AI tools, you may instead find it’s more beneficial to take an early draft and work on honing it yourself by hand. That allows you to put more of your original stamp on the roadmap that will take you through the remainder of the project.
Ultimately, that should create a better educational experience for you as well as a cleaner read for your audience.
Research
Research is considered a key piece of final projects. It’s a demonstration that you can do more than just take in information presented cookie-crumb style in a classroom. It validates your ability to seek out new information and incorporate it into an organized investigation. Your ability to develop a mastery of the tools and systems used to uncover information in computer science is considered key to your ongoing success.
Conducting literature review is one of the most time-consuming pieces of the final project process. Although there aren’t yet AI tools to fully automate this, it’s possible to loop through standard chat bots to speed up your research quite a bit.
There are also tools emerging like SciSpace or Elicit.org, which bundle the entire process into a single search query. Designed specifically for academic use, they offer quick prompts to narrow the focus and take your interests in the right direction. And since scientists and researchers in the field are already relying on them, it’s taking your understanding of AI uses in a direction the field is clearly heading.
Just as during the outline process, prompting an LLM to help unearth relevant papers can give you a good first pass at the articles you’ll need to go through. In fact, you probably already have a stack of references from your proposal.
By feeding the text of those articles or abstracts to the bot, you can ask it to analyze them and develop key themes in the pieces. You can tune those to your own topic as needed.
Finally, you can simply ask the bot to generate a complete literature review and sit back while it spits out a summary and citations.
Understanding How AI in Research and Review Can Be Useful Without Replacing Your Individual Inspiration
This is the phase where you start to get into some genuine questions about the utility of AI tools. Is it serving your ultimate purpose to have the AI read and summarize existing knowledge? In the best case, it can pull in and familiarize you with a much wider range of material than you could possibly review on your own. In the worst, it will, for vague and uncertain algorithmic reasons, miss some vital article that a human mind might instantly seize on.
What about doing the digging for source papers yourself but using AI tools to review and summarize the literature for you instead of reading every word. Isn’t it similar to skimming the paper for relevant points? Which, let’s be honest, is what you were going to do anyway after a hard night of gaming, right?
In the current moment, probably the best approach is to use AI in a way that actually increases your workload. That is, dive into your research to the best of your meager human abilities and turn up what you can… and then use AI to check on and expand your work.
Not only will you get the insights of your own efforts, but you’ll also absorb quite a lot of background information. The negative space – information that isn’t directly relevant, but points to what is – can do a lot to shape your thesis even if you don’t use it directly.
And you’ll open up the possibility that AI will turn up articles you didn’t find yourself, or highlight key information from others that you hadn’t noticed.
Writing and Revision
Thesis writing isn’t writing; it’s revising.
Pearson International’s 2020 Higher Education Learner Survey found that more than a third of college students aren’t confident of being able to identify problems in their academic writing. Even fewer felt like they could get support for polishing that work.
AI offers an amazing resource for students without strong English language skills. But it also begins here to touch on the part of the process that is most fraught.
Spelling and grammar checkers are already integrated into the most popular word processing programs that people use to produce thesis work. Those checkers are getting upgraded through AI bit-by-bit. Many now use autocomplete to guess the next word or phrase you likely intend to use. Even if you aren’t trying, you’re probably already taking advantage of AI to improve your writing to some extent.
Formatting is just the next natural step in that process. AI formatting tools help you structure your ideas to flow along a path that aligns with the exacting standards that compsci departments establish for submissions.
These tools, or close relatives, also ensure your citations stick to the required format. They may use AI-based plagiarism checkers as part of the package, ensuring you have no accidental inclusion of other people’s work.
Making the Big Decision About How Much GPT to Put into Your Thesis Paper
But the biggest question in the near term will be, how much should you rely on LLM-driven GPTs to actually form or revise the text of your thesis or capstone paper.
The academic writing skills of GPT models are pristine. With decades of exquisitely honed writing to tap into, they can instantly create papers that easily meet the technical standards of any journal you’d care to name.
The most difficult decisions you’ll have to make in using AI to help you with your thesis or capstone project will be in the writing phase.
As we’ve learned, with great power comes great responsibility. So you’ll have to tailor your use of AI in writing carefully to meet the goals of your project.
Addressing the Plagiarism Argument
A major goal of every thesis paper or capstone project in a master’s degree in AI is to develop original ideas within the field.
Yet it’s hard to hold out LLM based text generators as original by any measure. Rather than thinking machines, they represent statistical predictors of likely completions of the prompts they are given. While that process frequently results in non-duplicative statements that are clear and concise, they are still statistically likely segments of text as it exists within their corpus.
Much has been made of the possibility of these models apparently inventing facts or outright lies. But the very idea of invention misunderstands the system. LLMs can’t know when they are right any more than when they are wrong… they can’t, in fact, “know” anything in any human sense of the word.
Making matters more complicated for academic use of AI in thesis writing and capstone projects, AI that has been trained on more limited subsets of text — such as specialized academic fields where relatively few papers have been published — shows a markedly increased likelihood of it outright copying certain segments of text.
But AI is being shoehorned into university definitions of plagiarism in ways that don’t really respect the models, either.
According to the Cambridge Dictionary, plagiarism is “the process or practice of using another person’s ideas or work and pretending that it is your own.” Yet with AI, there is no person and no ideas in the sense that we understand either of those terms. The prompt, of course, comes from you; the words from an impenetrable algorithm that could only work from that prompt.
Naturally, your thesis work will have to comply with your university ethics rules. Many have now been contorted to something along the lines of “another source of ideas.” Yet the entire design of research and investigation in the culminating project is designed to spark ideas from other sources.
But these are exactly the issues you should be probing and defining as a graduate student of artificial intelligence. Get into the mix and get your professors on board as you explore a new frontier in academia.
Of course, these can vary from project to project. Ultimately, you want the text you develop to best express your own thoughts and concepts. But you also want it to communicate those thoughts as clearly as possible. To whatever extent AI can help you achieve that, it’s a net plus to both your education and to the state of progress in the field.
Final Step: Integrating Your Knowledge with AI Tools and Making the Final Presentation Your Own
No one blinks when you use a word processor with spellcheck to type up a thesis. Foreign students may rely on translation apps and it’s no big deal.
But when those machines start to approach some level of ability that begins to approach thought itself, suddenly academia is up in arms.
It’s the final part of the culminating project process, the presentation, where this becomes most problematic. Taken to the logical extreme, an entire credible thesis paper could come entirely from the mind of the machine… just using prompts, the existing literature and statistical models can produce a publishable thesis paper. A less-than-honorable student could follow the completions where they lead, learn the material, and successfully defend such a paper.
Odds are it has already happened somewhere.
But it’s not serving that student, their university, or the overall field of AI. Your presentation, or thesis defense, is where you fully demonstrate that you have developed your own thoughts on the project topic. And it’s one part of the project where AI can’t hold your hand.
Your Presentation or Defense Will Require Full Mastery of the Material
Even if you’re using AI in the best of faith, it’s easy to see how it can become a crutch. Rely on a summary without fully absorbing the source; let a GPT tool polish up your writing using a few words you don’t entirely understand. It’s a slippery slope.
You don’t want to be at the bottom of that hill when you step up in front of your thesis committee.
Thesis papers at the master’s level may require a full-on thesis defense. That’s you going in front of your committee and both orally presenting and discussing your paper.
Capstone presentations are also often delivered in a similar format. Given the more practical nature of capstone projects, though, it’s likely to take on more the nature of a demonstration. Technical details, like code presentations, will serve to show your expertise. More and more, these are delivered in video format.
Truth is found neither in the thesis nor the antithesis, but in an emergent synthesis which reconciles the two.
Some programs allow you to simply turn in your thesis paper and receive feedback and a grade (almost always pass/pass with revisions or fail). But an old-fashioned master’s thesis defense will put you on the spot. You’ll have to demonstrate a comprehensive understanding of the material and come up with acceptable answers to questions your committee has.
The only adequate preparation for this is to fully integrate your work with anything that you have used AI for. In that way, it’s sort of the ultimate expression of what AI should become. And that’s a valuable lesson for any grad student in artificial intelligence to take away from their studies.
As an AI Master’s Student, Your Understanding of how GPT Works Will Naturally Make You Cautious
One caution in using current generations of AI to help support your thesis or capstone project is that much of it is based on Generated Pre-Trained Transformer (GPT) Large Language Models. Fed on a diet of diverse written works drawn from published literature, the web, articles, and research papers, it’s thought to approach 600GB of text. That’s a ton of information when you consider the data density of the written word.
Yet it has a limitation: since it is pre-trained, all that data has a time stamp on it. In a field like AI where new breakthroughs are happening almost weekly, that means answers to your questions coming from ChatGPT and similar resources can be well out-of-date. And that’s not good enough for a graduate research project.
Another word of warning: since it is early days in the AI bonanza, there are plenty of charlatans mixed in with the prospectors trying to get in on the gold rush. You’ll run across tools that marketers have eagerly labeled as AI but in fact are built out of a handful of weak combinatorial algorithms and if-then logic that simply shuffles your stuff back to you in a slightly different order. Buyer beware!
But consider sorting through these charlatans as an important part of your education as an artificial intelligence professional.
Having said all this, it’s still also entirely common for a master’s thesis or capstone in AI to not involve supportive AI use at all, apart from spelling and grammar checks. Old-fashioned individual investigation, manual experimentation, and human-centered reasoning are still the standard. Since many thesis topics involve straightforward observation of performance or outputs, they aren’t necessarily about demonstrating technical skill in the programming or use of AI.
Instead, they are a good old-fashioned exercise in breaking apart problems facing the field and developing relevant data through regular investigation and standard data processing methods. You shouldn’t feel compelled to use AI in a novel way to complete your thesis just because it’s your topic of study. Use it where it makes sense to further your education.
Will Using AI Eventually Become the New Standard for Master’s Theses and Capstone Projects?
AI is a field that is evolving rapidly. You’re going to be part of pushing that accelerator pedal down even further.
So be prepared to make every piece of advice you’ve read here outdated.
Academia will come to terms with the abilities and tools you contribute to creating. As AI improves and becomes more commonly used in academic research, resistance to such tools may become a thing of the past. It’s possible that this perspective is as temporary as the way calculators were once viewed. Perhaps the abstraction of research skills is the next step in computer science. Maybe the AI tool is a feature that allows your mind to spend more time on integration and invention.
Finally, AI tools are coming out that take auto-complete to the next level. With a prompt as short as four or five words, systems can create several paragraphs of scientific text… with the citations to back it up.
It seems likely that, as with autocorrect, it will eventually be seen as more important that the results offer clarity and innovation… regardless of whether they were typed by man or machine.
While most of the issues discussed so far seem rooted in Generative Pretrained Transformers (GPT), it’s worth highlighting that these are only one kind of AI. If they are currently the most relevant hotspot for textual use that happens to line up with thesis and capstone project writing, they’re not the end of the story.
Sufficiently advanced AI may become a sort of collaborator on thesis projects. Just like you name off all the students who participated in a capstone, you might eventually be putting down the name and nomenclature of the latest AI tool as your co-author.
Expecting the Unexpected: How AI May Make Thesis Work Even Tougher
The wider use of AI in master’s thesis and capstone projects might also come with a less obvious effect: raising the bar for those projects.
There’s no reason to think that AI automation in graduate studies will have any different effect than we expect in other kinds of work. If you can accomplish more, more perfectly, in less time, expectations will rise.
For example, it’s no longer remotely acceptable to have typos in thesis papers in a world where spell-checkers exist. As AI becomes more common, there will be less and less leeway for imperfection in other parts of the project. Glitches in experiment design won’t be tolerated where AI can put together a perfect set-up. Missing a relevant paper in literature review can’t be overlooked when AI surely would have found it.
You may be expected to take on bigger questions, find more answers, expand the field further when you are relying on AI to support your project. Just as it’s expected to be a tool to multiply productivity in business, similar results should be seen in the academic world. Otherwise, why bother creating it?
Finding a Place for Human Serendipity May Be the Ultimate Challenge When Using AI
Yet there’s also room to question whether such perfection is ultimately in service of better science. You can’t look at the history of scientific breakthroughs without noting how many were the result of various kinds of experimental glitches. From penicillin to safety glasses to the cardiac pacemaker, plenty of important concepts and inventions have come from human error.
Does AI leave room for serendipity in your thesis or final project?
And what about the random strike of inspiration that comes from mistakenly going down the wrong path? If AI smooths out your research process so that you always get where you are heading, what happens to the mistakes that change the world… radar studies that lead to microwave ovens, a chamber for studying cloud formation that ended up unlocking the secrets of nuclear energy?
Someone should write a thesis about it.