Can AI Write Academic Papers? 5 Key Things to Assess

1. What is AI?

2. How does AI work?
3. AI writing tools

4. What can AI writing tools do?
5. Can AI write academic papers?
Concluding remark

This article discusses artificial intelligence (AI), its roles in various industries, its potential in the future, and its capability to write academic papers.

Being in academic publishing for more than a decade, I am intrigued by what the future holds for AI in scholarly writing and academia. And I always want to know, can AI write academic papers?

Previously, I wrote on different ways publishers use AI, but this write-up zooms more detail into AI and academic writing.

Of recent, I have been exposed to several AI writing tools and have been able to see hands-on how the tools work, the strengths, potential and limitations, with a particular focus on academic writing.

This article will not discuss specific AI writing tools but rather the general features and their potential in the writing arena.

Could one day the AI will take over a researcher’s role in the academic community? If the answer is yes, to what extent?

Will AI be able to write high-quality papers, grants and proposals on demand?

1. What is AI?

AI is a system created to emulate cognitive functions that are specific to humans. The simplest way to understand AI is to think of it as a computer that learns and makes decisions. As it learns, it gets better.

A computer system can perform human intelligence tasks with AI, such as visual perception, speech recognition, translating, decision-making, risk-taking, and prompting motions.

The term “artificial intelligence” was first coined in 1955 (the year Steve Jobs and Bill Gates were born) by John McCarthy, a computer science professor at Stanford University, United States. He wrote a proposal to organize a small conference on artificial intelligence at Dartmouth College, New Hampshire.

Since then, AI has grown by leaps and bounds, finding its place in a broad field that focuses on developing intelligent machines for specific tasks. Some of the tasks include voice recognition, decision-making, facial recognition, language translation, among others.

In 1997, an AI machine called Deep Blue developed by IBM scientists defeated Garry Kasparov, undefeated world champion, in a chess game. The highly-publicized win established AI as a force to be reckoned with.

AI continues to be widely used in many industries today: automotive, engineering, IT, medical, military, education, data processing, security and many others.

AI is also being used in the creation of autonomous vehicles, unmanned aerial vehicles (UAVs), drones and other advanced machines that will soon be smarter and safer than handled by humans.

2. How does AI work?

AI works based on data inputs. The more inputs the system is fed, the better decisions/outputs it will produce. Good AI systems, therefore, works based on exabytes of data.

Driven by data, AI systems can perform complex tasks that seem intelligent – from translating languages to driving cars.

AI can also solve problems and find specific information, such as answering a question like “how many countries are in Africa?” AI systems must be able to correctly interpret data to give accurate results.

While AI systems can be very accurate in decisions, there are times when it fails too. This is because AI systems work based on the information they are fed, which may or may not be accurate.

For instance, an AI system based on data from a world map created in the 1800s will perform very poorly. This is because the world’s geography has changed significantly since then.

AI systems also face the “garbage in, garbage out” problem. If the information fed into the system is inaccurate, then the output will be wrong too.

3. AI writing tools

AI-based writing tools (also typically called “AI writers” or in this article simply AI writing tools) are software that use AI technology to produce texts. Using an AI writing tool, the system will write content (the output) based on your instruction (the input).

But how does the AI system know the kind of output to produce?

A good AI writing tool needs to be fed with tons of information; in most cases, they were fed with information on the Internet and the public domain. The more the details are fed, the better the AI can learn, and the better the output.

AI writers can emulate the way humans write and “think.” The use of AI writers is gaining momentum as they provide several benefits to academic and non-academic writers. One of the main advantages is that it speeds up writing.

From finding information to structuring, the whole process becomes faster and more efficient than when a human writes.

4. What can AI writing tools do?

Now, here is an interesting part. AI writing tools can do many things depending on what you tell them to do. Here are just some of the things AI writing tools can do:

Write texts, including captions, social media posts, letters, emails, essays, articles, or news reports.

Check for grammar and spelling errors using natural language processing (NLP) technologies. NLP is a type of AI that gives machines the ability to read and understand human language.
Synchronize different documents to ensure consistency among them.
Format text into any citation style you want (e.g., APA, MLA)

Having worked with several AI writing tools, I found that some offer fantastic features highly attractive to writers, copywriters, and content creators.

The subsequent section will discuss some of the capabilities of AI writing tools.

4.1. AI can write short and long-form content

AI writing tools can write short copywriting content (for social media posts, captions, etc.) and long-form content (blog write-up, articles, letters, etc.).

In writing and content creation, producing content of different lengths and structures is essential because different content is used for different purposes. For instance, social media posts are usually much shorter than regular articles.

A good AI writing tool will produce quality content in any given length or structure that it has been fed with details.

AI writing tools are not limited to a particular content genre, be it persuasive or informative texts. This means that AI writing tools have the potential to disrupt the academic publishing industry.

AI writing tools can work based on templates, but it is even more interesting to follow dynamic templates. The templates are designed to include specific keywords based on your preferred style and tone.

This means that every time you use the tool, the output will be different. This provides flexibility that makes these AI writing tools highly attractive for content marketers.

4.2. AI can write fast

Another promising feature of the AI writing tool is its ability to produce content fast. Producing content at a blazing speed is especially useful in high-workload tasks that require writing several pieces of content within a short time.

The advancement of academics will also largely rely on their publication productivity. The more they publish, the faster they can climb the growth ladder. Hence, anything that helps them write academic papers fast is a god-sent.

Due to the nature of content size they have to work on daily, big content creators from other industries will find this ability appealing.

4.3. AI can help you find information fast

Suppose that you are looking for the answer to a specific question, such as the health benefits of drinking water. AI writing tools can quickly search through thousands of documents and provide the answer within seconds.

This feature is very helpful as it can save you time that would otherwise be spent doing the legwork yourself. And it is especially useful in research and academic writing where time is extremely valuable.

4.4. AI can be integrated with other tools

The AI writing tool will integrate with other tools, such as Grammarly for language checking, Copyscape for plagiarism check, and search engine optimization (SEO) tools.

More advanced AI writing systems also come with API integration, in which you can use the tool within another system, such as inside a Google document and email client.

Integrating these AI writing tools means that you have a powerful writing arsenal at your disposal that can provide the best content.

4.5. You can instruct the AI to use a certain tone

With the AI writer, you can also influence the system to write in a certain way or tone by providing relevant instruction. An article can be written in a casual, formal or even witty according to your need.

For example, if you are writing for a mainstream newspaper or magazine, you want your article to be formal or serious. If you are writing a blog post, you can tune it down or make it more interesting using a witty tone.

The ability to influence the writing output means you can use AI to customize to suit prevalent needs. Cool, huh?

4.6. AI can produce plagiarism-free content

Other than integrating with plagiarism tools, AI writing tools integrate their own built-in plagiarism system. You can be ensured of plagiarism-free content using this system.

5. Can AI write academic papers?

Writing academic papers require many steps. And a research publication such as a journal article needs to possess a few characteristics to be called an “academic work.”

Some of these steps, such as analysis and synthesis of research data, require a high level of thinking. In fact, they cannot be performed by modern tools, including AI, as yet.

Some other challenges that AI faces in writing academic papers will be as follows.

5.1. Original research

In writing academic or research papers, you will need to identify a research gap not covered in the literature. In other words, you need to produce original research work. Otherwise, your academic papers will be discredited and bring no merit in academia.

AI can write original, plagiarism-free texts and proses, but can they write original research? Not at this moment. Original research and unique content are two different things.

AI also will have issues accessing research behind a paywall, especially important journal articles requiring access fees. It can read abstracts but not the entire article unless the journal publishes under the open access model.

5.2. Interpret and challenge research from literature

AI tool can “understand” what is written in the literature and rewrite it, but it has yet to reach a level to interpret complex analysis, let alone challenge the findings of other studies.

So far, AI tools are good in writing extended summaries of the already published content. But they cannot write original research that challenges the findings of another study.

5.3. AI does not know the recent events

This is one of the most obvious pitfalls of AI writing tools now.

None of the AI tools I am working with have been fed with information for 2020 and beyond.

In other words, the AI cannot tell you much about the Covid-19 pandemic that began in late 2019; you need to feed it with plenty of information for it to digest and churn out content.

5.4. AI writing tools are more of a writing assistant, not a writer per se

One common remark I noticed made by quite many people is that AI writers often produce gibberish content.

At a closer examination, I discovered that they did not feed the right input into the system; rather, they ran a command and hoped for a miraculous output, such as a 1,000-word article with just a click of a button.

You must think of the AI writing tool as an assistant, someone who gives you that nudge when faced with writer’s block issues.

With the AI writing tools, you can generate thousands of ideas within minutes by simply feeding your brain what you are writing and other sources. Sometimes, by feeding random sources, the AI can also help you discover important outlines that have not crossed your mind.

Do not treat AI as a system that writes you a complete article in a single command without doing your part in giving the right input and guiding the AI system.

5.5. AI cannot fact-check reliably

AI writing tools today are still lacking in carrying out fact-checking reliably. They cannot accurately interpret data abstracts and other contents in the literature.

In factual work, AI writing tools can give you a pointer if the content has been copied from somewhere else or not. This is why many of them integrate plagiarism-checker before feeding into their word processing system.

Currently, when using AI writing tools, you will need to ensure that facts, numbers and statistics undergo thorough fact-checking and reviews to avoid embarrassment.

5.6. Censorship in AI writing tools

Most AI writing tools have some sort of censorship or content filter when writing content. For instance, when I commanded the AI to write about the Khmer Rouge’s brutal policies when it took over Cambodia in 1975, it refused to write and responded with a label saying “unsafe.”

If you are a historian writing about certain unpleasant historical accounts, you may end up with troubles getting your writing done with AI.

A few of my acquaintances have also confirmed that they cannot prompt the AI systems to write adult content.

Concluding remark

AI still has some way to fully write academic papers. Nonetheless, it is doing a sterling job to continue transforming the way we do things and changing the publishing and writing industry.

Each AI writing tool offers different features and experiences. The pricing also varies largely depending on the functionalities, complexities and levels of customization.

The best way to know which one would work for you is to try using them yourselves. Keep also in mind that, in publishing our research work, we do not just publish papers; we want to write good scholarly articles that fulfill vigorous research and academic standards.

The advancement of AI writing tools has also triggered discussions on ethics and principles.

Some academics and scientists worry that the widespread use of AI in academia could adversely affect students’ education and promote “laziness.” Another worry also AI produces academic papers by simply looking for patterns in text data sets rather than understanding the meanings behind data sets or ideas.

AI is still capable of writing good stuff for academics and researchers, albeit with limited capabilities. AI writing tools will also help researchers to be more productive and work faster.

Freeing up time from the writing responsibility may open up more opportunities for an academic to make new scientific discoveries that may change their field of study.

No doubt, AI writing tools will continue gaining popularity, and their use will become widespread in the research and academia as of this writing.