I have written a few blogposts about the problems with using AI for writing. This time I asked ChatGPT to list 10 reasons why science writers should not use AI to write their papers. I thought it would be interesting to see how/if AI criticises itself. The following is verbatim:

My prompt: Can you give me 10 reasons why science writers should not use AI to write their papers?

"Yes. If your audience is scientists, it's important to make nuanced arguments. AI is a useful tool for some tasks, but there are strong reasons not to rely on it to write a research paper. Here are ten reasons you could use in a workshop, article or discussion.

1. AI cannot take responsibility for the research
AI did not conduct the experiments, collect the data or make the scientific decisions. The authors are responsible for every claim in the paper, regardless of who or what drafted the text.

2. AI can invent information
Large language models sometimes generate incorrect facts, citations, statistical interpretations or methodological details that sound convincing. These errors can be difficult to detect, particularly outside the author's area of expertise.

3. AI can weaken scientific thinking
Writing is part of the research process. As scientists write, they clarify their arguments, identify gaps in their reasoning and refine their interpretation of the results. Outsourcing the writing can also outsource some of this thinking.

4. AI often produces generic writing
Scientific papers need to communicate the novelty and significance of the work. AI tends to produce safe, conventional language that can obscure what is genuinely original about the research.

5. The writing may no longer reflect the authors' understanding
A paper should represent what the authors know and can defend. If AI introduces concepts, interpretations or terminology the authors do not fully understand, they may struggle to answer reviewers' questions or explain their conclusions.

6. AI can introduce subtle inaccuracies
Even when the overall text appears correct, AI may change the meaning of a sentence, oversimplify complex ideas or overstate the certainty of findings. These changes can alter the scientific message.

9. Scientists need to develop their own writing skills
Clear writing is an essential research skill. It influences publication success, grant applications, collaboration and career progression. Relying heavily on AI may slow the development of these skills, particularly for early-career researchers.

10. Good scientific writing requires judgement
Choosing what to include, what to omit, how to organise evidence and how strongly to interpret results all require expert judgement. AI can imitate these decisions but cannot replace the scientific expertise behind them.

A balanced perspective
Rather than saying scientists should never use AI, a more defensible position is:
- Don't use AI to generate the scientific content, interpretation or conclusions of your paper.
- Do consider using AI for low-risk tasks such as brainstorming titles, checking grammar, suggesting alternative wording, summarising your own notes, or identifying awkward sentences—provided you verify every suggestion and comply with your journal's policies.

This distinction encourages scientists to retain ownership of the intellectual work while using AI as an editing or productivity aid rather than a substitute for scientific writing."

While I agree that it's ok to use AI for 'low risk tasks', I do not include summarising in this category. Whatever you use AI for, consider it the first author or a co-author if you use it's exact wording or summary points.

© Dr Marina Hurley 2026 www.writingclearscience.com.au

Any suggestions or comments please email admin@writingclearscience.com.au 

Find out more about my online courses...

SUBSCRIBE to the Writing Clear Science Newsletter

to keep informed about our latest blogs, webinars and writing courses.

There are many obstacles that can prevent us from being productive and efficient writers, especially procrastination. Even if there is a looming deadline and we are well aware of what needs to be written, sometimes we simply don’t feel like writing. And if we don’t feel like writing, procrastination can creep in, and we might put off writing until the last minute and then produce something that is substandard or incomplete.

Here are four ways to set you on the path of writing, even when you don’t feel like it.

1. Write intensively for short blocks of time

Set a timer and commit to writing for only 30 minutes. Thirty minutes of writing might seem achievable when you don’t feel like writing. Tell yourself that after the 30 minutes is up, as a reward, you will allow yourself to do whatever you want for one hour. Commit to only write for 30 minutes and make sure that you don’t do anything else during that period: no re-reading what you wrote previously, no stopping to google something, no telephone, no talking. Just write. If you find this too hard, start with ten minutes.

If you are normally a productive writer, 30 minutes might not seem like a long time, yet 30 minutes writing is much better than not writing at all. We can write a lot within short time periods if we don’t allow ourselves to get distracted, especially with other work tasks.

Avoid setting the timer for too long a period. Avoid setting unrealistic goals for yourself as you will feel unproductive and unsatisfied if you don’t meet them.

When you set a timer, try placing it out of reach so you have to get up from your computer to turn it off. After 30 minutes, stand up and walk around, print out what you have written so you have physical evidence of your productivity. You may find that you want to reset the timer for another 30 minutes and keep going. Ultimately, you may find that setting these short writing sessions allows you to break through the barrier of just getting started.

2. Don’t switch between writing tasks during a designated writing session

During your writing sessions, only compete one type of writing. Switching between different writing tasks will sap your energy more quickly than completing one task. If you are getting your thoughts down, just write freely and don’t switch to editing or proofreading halfway through. If you are rewriting a section of your document, don’t switch to writing new material on a related topic. 

3. Take a break from the computer: print your document out and use a pen

We spend a large proportion of our time in front of a computer for all sorts of work activities, especially writing. Periodically take a break from writing on a computer, even if it is for a short time. Try printing out the latest draft of your document and take it to a café or a lounge chair with a pen and a notebook. Edit your draft by hand and use a notebook to write fresh material. You might find that the change in environment allows you to relax yet you can still work on your document.

I always print my document out whenever I complete a draft so that I can see how my writing looks on paper and then plan what writing I will do for the next draft. With a paper printout, I can see my whole document at a glance, without having to scroll through a digital version on a computer screen. Moving to a lounge, or a comfortable environment, gives me a break from staring at a computer screen and sitting on an office chair.

4. Find a friend to write with

Writing is usually a solitary activity but you might it more enjoyable if you organise writing sessions with a friend or colleague. Try taking your laptops to a park or a café and set up 30-minute writing sessions followed by 30 minutes social chat. You could also read and give feedback on each other’s work.

And remember…

Avoid creating unrealistic expectations that create stress and reduce our work satisfaction. It is unrealistic to expect that we can be super-productive writing machines that can write anything, anywhere and at any time. Aim to write in intensive pre-organised, short blocks of time in an environment that is as comfortable and distraction-free as possible.

© Dr Marina Hurley 2026 www.writingclearscience.com.au

FURTHER READING

• 11 common mistakes when writing an abstract
• The essentials of science writing: identify your target audience
• Co-authors should define their roles and responsibilities before they start writing
• Critique all writing advice
• When to cite and when not to
  
• Back to basics: science knowledge is gained while information is produced
  
• How to build and maintain confidence as a writer
• If science was perfect, it wouldn’t be science
• The essentials of science writing: What is science writing?
• 8 steps to writing your first draft
• Two ways to be an inefficient writer
• Work-procrastination: important stuff that keeps us from writing

Any suggestions or comments please email admin@writingclearscience.com.au 

Find out more about our online courses...

SUBSCRIBE to the Writing Clear Science Newsletter

to keep informed about our latest blogs, webinars and writing courses.

Applying for grants is a time-consuming process. Deadlines can loom suddenly, leading to stress and long days. Success rates can be low, meaning that considerable effort is often not rewarded. 

At the outset, you need to form a realistic timeline to work on the grant application. In addition to identifying your funding source, researching the specific requirements of the grant application and formulating your project budget, you will need to start planning your written proposal. Careful planning and early feedback from your colleagues will maximise your chances of a successful outcome. 

In addition to the detailed advice you will receive from your grant body, following the 6 steps outlined in this article will help you develop the core content of your application. 

Step 1: Identify your audience

Will the grant be reviewed by specialists in your field and will the panel include non-specialists or lay people? In reality, you may have two types of audience:

- Non-specialist or non-scientific reviewers who will require background information to judge whether your project is worthwhile.

- Specialist reviewers within your discipline that will be more familiar with your research.

Depending upon the funding organisation, some of your reviewers may not be familiar with the terminology or the current research problems of your discipline or immediately appreciate why your project is so important. Therefore, it is imperative to make sure your proposal outlines the relevancy of your topic to current issues and that it is clear and concise.

Step 2: Summarise the key research problems overarching your project

Before you can think about convincing people how good your project is, you need to build a picture about the current problems facing your research community.

Successful grant applications clearly define the area of need and how it is relevant to your potential funding source. In addition to outlining the scientific importance of your project, describe how your work might affect society, the environment and the economy. For example, if you are researching a disease, highlight the burden associated with that disease. Outline how many people it affects and the costs to society.

Step 3: Summarise the key problem your project will solve

This is where you focus on the key problem your project will attempt to solve. Clearly articulate what specific problem needs to be investigated and why. Avoid promising to solve too many problems or problems that are too large in scope, otherwise your project will seem unrealistic and unachievable. Describe how the key problem is connected to the broader scope of the research problems outlined in Step 2. Once you have outlined your research problem, then you can clearly state what you aim to achieve (step 4).

Avoid writing vague or generalised statements, for example: “To date, little research has been conducted in this area.” To help you keep your application short, also avoid obvious statements, for example: “This topic was investigated through an extensive search of the literature.”

Step 4: Articulate your aim, hypotheses and outcomes

While directly referring to your key research problem, outline what you aim to achieve, including your hypotheses and what outcomes you can expect from your completed project.

Once the overall aim is stated, the project should be broken down into sub-aims, each with a defined outcome. This approach helps you to define realistic timelines, ensures that the project description is concise and will improve the likelihood that the grant will be successful.

Step 5: Summarise how you will do the work (methods)

A major factor in grant success is being able to convince the reviewers that the project is feasible and important and that the work is likely to be completed within the proposed timeframe. Clearly outline what methods you will use and what experience you have in this area. If you need to develop new methods, clearly explain what is required and provide evidence of your ability to develop other methods in the past.

Outline the scope of the project. How long it will take to complete each component? Is the size of your project feasible within the set time frame? Do you have access to suitable equipment and operational facilities?  Promote yourself. Provide evidence (such as previous publications or unpublished data) to demonstrate that you are capable of successfully completing the project.

6. Seek feedback from colleagues

Give your draft proposal to your colleagues for feedback. They may provide valuable feedback on what is feasible, which aspects are the most interesting and what might be missing.

This early feedback will help you focus on what you want to achieve, why it is important and how likely the project is to succeed. It can be helpful to talk to people who have already received funding from a particular source; what feedback did they receive and what aspects did they think helped them to secure funding? If appropriate, it might also be helpful to seek feedback from colleagues who have recently been unsuccessful in winning a grant from the same funding body.

What to do next?

- After receiving feedback from your colleagues, rework the application so that it is clear, compelling, concise and flows well.

- Finalise your budget and ensure all aspects of your project are justified.

- Seek at least two more rounds of feedback from your peers as you proceed through the writing and submission process. Grants that are peer-reviewed prior to submission are more likely to be successful.

- Pay close attention to the small details in the submission process. You don’t want to have your grant rejected on a technicality or an unchecked box on a submission form.

© Marina Hurley 2026  www.writingclearscience.com.au

Any suggestions or comments please email admin@writingclearscience.com.au 

FURTHER READING

• Should we use active or passive voice?
• 10 writing tips for the struggling ESL science writer
• Co-authors should define their roles and responsibilities before they start writing
• How to write when you don’t feel like it
• When to cite and when not to
• Back to basics: science knowledge is gained while information is produced
  
• How to build and maintain confidence as a writer
• If science was perfect, it wouldn’t be science
• The essentials of science writing: What is science writing?
• 8 steps to writing your first draft
• Two ways to be an inefficient writer
• Work-procrastination: important stuff that keeps us from writing

Find out more about our online courses...

SUBSCRIBE to the Writing Clear Science Newsletter

to keep informed about our latest blogs, webinars and writing courses.