Pages

Wednesday, 17 June 2026

The ethics of LLMs

AI seems to be becoming more and more pervasive, but really 'AI' is the increased use of large language models, or LLMs, as a writing and communication tool. LLMs such as ChatGPT, Gemini, and Microsoft Copilot are becoming a standard tool in workplaces. Since the Covid-19 pandemic, it seems that employees are being required to more consistently use digital skills, and - when seeking work - to show evidence of how digitally-able they are. 

That means that we, as career practitioners, must also be familiar with the skill sets our clients are likely to need in the workplace. As practitioners we also face some ethical dilemmas: firstly, about responsible AI tools use; and secondly how to support clients using these tools (Sedeeq & LaFever, 2025).

While younger workers seem less worried about adopting LLM tools, and less worried about their use; older workers worry more about who actually owns the words we are pulling from the ether, and what the End User Licence Agreement (EULA) says. Meanwhile the research shows that regular use of LLMs negatively affects our brain development (Kosmyna et al., 2025); and there are institutional disputes over LLM tool authenticity, security, and accountability (Sedeeq & LaFever, 2025). The question arises: who owns these LLM-generated products?

Further, if we - or our clients - are studying and use the LLM analysis and synthesis tools, does that actually count as 'genuine learning' (Sedeeq & LaFever, 2025)? A Pew study reported "13% of 13-17 years olds had used ChatGPT to complete schoolwork", growing by early 2025 to 26% (Sedeeq & LaFever, 2025, p. 22, citing Sidoti & Gottfried, 2023). How does outsourcing of education sit with us ethically?

Then we have bias, which is apparently endemic within LLMs (Sedeeq & LaFever, 2025) simply due to the WEIRD pool of material sourced to underpin it: Western, educated, industrialised, rich and democratic (Henrich et al., 2010; read more here). Oh, and culturally? Mainly North American and middle class (Sedeeq & LaFever, 2025). So how does that sit with cultural diversity within our practice?

Sedeeq and LaFever point out that "Using AI as a complementary tool such as to identify trends in a dataset or simplify language in a difficult text is an appropriate extension of traditional learning aids like a dictionary", but warn that "if [our] clients are relying on AI to generate content that they submit as their own with limited to no modification, they have crossed an ethical line" (2025, p. 23). But is that our problem as practitioners?

I guess it is if we are doing it. I am less certain it is our gig, if it is our clients who are doing it. 

We need to think about the intent of the LLM ownership. Anthropic, owner of ChatGPT guided the missiles and drones bombing Iran (Bregman, 2026): should we use a tool which is used to attack sovereign nations?

So, to summarise: who owns the words; can education be outsourced; is cultural homogeneity what we want; can we contract out our work; and who owns the model and what is their intent? I am sure there will be many more questions which need answering, but this is starting to look like a bit of a can of worms.

Actually, this is starting to remind me of Developer Bob (Young, 2013).


Sam

References:

Bregman, R. (2026, March 4). Quit ChatGPT: right now! Your subscription is bankrolling authoritarianism. The Guardian. https://www.theguardian.com/commentisfree/2026/mar/04/quit-chatgpt-subscription-boycott-silicon-valley

Henrich, J., Heine, S. J., & Norenzayan, A. (2010). The weirdest people in the world?. Behavioral and Brain Sciences, 33(2-3), 61-83. https://doi.org/10.1017/S0140525X0999152X

Kosmyna, N., Hauptmann, E., Yuan, Y. T., Situ, J., Liao, X. H., Beresnitzky, A. V., ... & Maes, P. (2025). Your brain on ChatGPT: Accumulation of cognitive debt when using an AI assistant for essay writing task. arXiv. Advance online publication. https://arxiv.org/abs/2506.08872

Sedeeq, R., & LaFever, C. (2025). Balancing Innovation and Integrity: Ethical Considerations for Using AI in Career Development. Career Developments, 41(2), 22-23. https://associationdatabase.com/aws/NCDA/asset_manager/get_file/436612?ver=1426&tcs-token=c5d672670ffda7c9af250c94151f65ab7cc8f66578ef1b7f03a4fc33f7a58f82

Young, D. (2013, February 20). Developer Bob - Outsourcing Your Job. https://www.samyoung.co.nz/2013/02/developer-bob-outsourcing-your-job.html

No comments :

Post a Comment

Thanks for your feedback. The elves will post it shortly.