Read Law Right

I’ve used four posts now to build an argument for taking care around your use of gen-AI. But I’ve tried to be careful not to completely malign it or cast it as technology’s forbidden fruit. So after all that:

Is there a place for gen-AI in your studies?

Maybe. But I recognise that the boundary line can be difficult to discern. It is easy enough in the abstract to say don’t use it to replace thinking, or for demonstrating understanding, analysis and evaluation.

The advice that follows absolutely must be filtered through whatever your university, institution, employer or professional regulator prescribes (or even proscribes) about using gen-AI. My thesis here is limited to the following: how can you harness the undoubtedly impressive power of this tech without outsourcing your agency.

Here are some principles and tests that I have developed alongside my own use of this technology.

An initial rule-of-thumb that can be expressed both positively and negatively:

• don’t prompt gen-AI with material its response to which you are badly placed to evaluate.
• consider prompting gen-AI only if you are well placed to evaluate what comes back.

Then, if you are still wondering whether to prompt gen-AI for a response, ask yourself:

• is this work intended to contribute to my intellectual development?
• is this a task on which my professional identity and reputation depend?
• is this a task where it is necessary to protect and exercise my agency?

And I suggest that in both the student and maybe even professional realms, you protect those values if you limit your use of gen-AI to work that:

• you have already researched and drafted yourself,
• to a reasonable standard that you are otherwise content with,
• where you can legitimately claim to be well placed to evaluate the AI’s reply precisely because of your originating work, and
• would be comfortable explaining how you used it to a professor, colleague or client.

Applying this to real life

Let’s make this even more real by applying it to how you would demonstrate understanding, analysis or evaluation.

You ask the AI to assess your work for clarity, and the AI makes suggestions that strengthen your explanation, or deepen your analysis or focus your evaluation:

Because you have done the reading and research comprising your prompt to the AI, you are in a strong position to evaluate the AI response. The AI response causes you to re-think, but it is still your thinking, your agency, at work when you decide on the edits.

The AI suggests a new case or source to improve your work:

Again, you are in a good place to then locate that source, read it, and evaluate its relevance to your work. That’s exactly what you would do if your professor or tutor suggested additional reading. The decision how to use that is yours, as should be the words you use to do so. You don’t then ask the AI to summarise and evaluate the suggested source for you!

The AI tells you that your understanding is wrong, your analysis is incomplete, or your evaluation is unpersuasive:

This is harder, but you still have the foundation in your own research and writing and now you have questions that you can follow up. Treat the AI response as its version of prompting you, effectively turning the tables back on you. Verify your understanding with the textbook. Check your analysis by returning to your sources: have you articulated the arguments accurately? Review your conclusions: do they flow naturally from the analysis?

In each case, you are the one carrying forward the thinking. You are the one improving the analysis. You are the one critically reassessing your evaluation and conclusions.

The trickier example: when the AI suggests a compelling form of words

This is the hardest example, but it cuts to the core of what gen-AI does. We move from AI commenting to AI contributing, which is qualitatively different. It gets much closer to some of the dangers highlighted in the earlier articles – that you switch off your own thinking and delegate to the machine. I don’t have an answer here that I’m entirely satisfied with, only principles and suggestions to help you think it through.

Let me be clear, I am not considering here an example where you instruct the AI to “Rewrite this draft to make it more academic” or anything approaching an AI-generated rewrite of your work. That would amount to a clear outsourcing of your agency. It may also amount to an academic offence depending on the rules of your institution.

Instead, I have in mind the sort of prompt, like those envisaged above, where you ask the AI for specific feedback on an aspect of your draft. And in reflecting back to you its understanding of your writing, the AI uses a specific form of words that you find compelling.

Let’s think about this in three different ways: how it affects intellectual agency, the extent to which authorship and attribution are in play, and how it intersects with the rules of academic integrity.

Intellectual agency

I have wrestled with these exact issues in the writing of this series. This is the first major block of writing on the blog in several months and the first where I have used gen-AI to help improve the quality of the writing.

The principle that I have alighted upon for myself is this. When I supply the AI with material that I drafted myself, I consider myself well-placed to evaluate the AI response. Therefore, when it produces a form of words that I prefer, I have sometimes adopted them.

I have done so satisfied that I have not thereby outsourced my thinking, professional judgement, intellectual development or agency.

Authorship/attribution

Then comes the question, if I use a word, phrase or sentence that the AI coined, whose words are they really?

The principle I have observed here is quite simple: I operate a clear distinction between the AI commenting on my own intellectual work, and the AI supplying substantive new material that I did not previously possess.

So if the AI coins a useful phrase out of my own original work that I adopt, I do not attribute that to the AI.

However, my work is not submitted for assessment or credits. Assessed work requires you to add a third and decisive consideration: academic integrity rules.

Academic integrity

And so my final piece of advice in this section legitimately applies to you as students in a way that it does not when merely authoring a blog post.

In any writing that you submit to your academic institution for marking and credits, you absolutely must follow their rules of academic integrity to ensure authorship is correctly attributed. If they say every word from any external source must be attributed to that source, you attribute it. Nothing in this article removes your obligation and responsibility for following those rules.

And remember, you can always reach out and ask the professor or tutor who will be marking your work for a definitive answer.

Epilogue

The work of study shows up in your understanding, analysis and evaluation, and your assignments are intended and designed for you to develop and demonstrate those objectives. But part of the task of understanding, analysis and evaluation pre-dates our demonstration of those things in writing and relates to the tools we use to develop them in the first place: a textbook; an article; a case; a discussion with the tutor or professor or fellow student, or gen-AI?

And that extends to evaluating the appropriateness of the tool to the job in hand. As an undergraduate student, I wouldn’t initially direct you to complicated journal articles that assume a degree of knowledge (or even a degree!) I’d start you off in the textbook. Evaluating the right tool for the job. We must treat gen-AI the same. We must ask, is this the appropriate tool for this job, now, given my current state of knowledge and understanding and progress on the problem. And always asking, have I done the intellectual work to equip myself to evaluate what it tells me, without sacrificing any of those values we talked about – most urgently, without sacrificing my agency?

If the answer to those questions is yes, it is undoubtedly a powerfully useful technology. The technology is new: the intellectual discipline required to use it isn’t.

If you’ve found this series helpful, subscribe below to get notifications when subsequent articles drop. The next season will be explaining the grading criteria and how to write in a way to satisfy the grading criteria. Meanwhile, check out seasons 1 and 2 on the reading and note-taking processes.

---

Enjoyed this article?

Read Law Right is a blog about reading effectively, writing persuasively, and developing the understanding, analysis and evaluation skills that help law students to succeed.

Subscribe below to receive notifications of future posts.