AI Disclaimer — July 2025

This disclaimer declares the ways in which I use generative artificial intelligence in my work, studies, and personal life. It is up to date as of July 2025. It covers products I pay for as well as free products, and is intended both as an example of what an AI disclaimer might look like in an academic context and as a reference I can point to whenever people ask how I use the technology.

1  Subscriptions

1.1  Paid subscriptions

ChatGPT Plus
Claude (paid tier)
Google One (AI features)
Zoom Business (AI features)
Midjourney
Microsoft Copilot (part of the Microsoft 365 subscription)
Canva (paid plan, including built‑in AI features)
Otter.ai

1.2  Free subscriptions

Perplexity (free tier)

1.3  Lapsed subscriptions

Perplexity Pro
Runway
ElevenLabs (professional voice‑clone subscription)

2  Work

The following non‑exhaustive list covers how I use GenAI in my consulting, professional development, and writing:

ChatGPT – Deep Research for blog post research tasks (e.g. comparing the Australian, UK, and US curricula for performance criteria related to digital technologies and sustainability, as used when creating overviews for posts like Teaching AI Ethics).
Otter.ai – transcribes audio recordings of draft articles.
ChatGPT o3 or Claude 4 Sonnet – correct transcription errors and format audio transcriptions for manual editing on the blog.
ChatGPT o3 – clarification and refinement of complex ideas (e.g. structural edits for long articles like AI and the Death of Communication).
ChatGPT‑4o (with internet search) – ad‑hoc research of articles related to blog posts and professional‑development sessions.
ChatGPT (any model) – creating PowerPoint templates (with code) for manual editing.
ChatGPT, Claude, Google Gemini, Microsoft Copilot – featured in instructional videos about these platforms for professional development sessions.
Claude 4 Sonnet – writing Python scripts for automations (e.g. blog scrapers to archive, back‑up, and work with blog‑post content directly from my website).

3  Studies

The following uses apply to my PhD studies. (Because participant data are anonymous, I do not use generative AI for data analysis: see section 4 below.)

Otter.ai + ChatGPT / Claude – recording and transcribing articles and thesis sections.
Claude 4 Sonnet and ChatGPT o3 – structural edits plus spelling, punctuation, and grammar (as declared in academic articles such as this one).
Elicit – occasional additions to literature reviews, especially when search queries are complex and traditional search (e.g. via EBSCO) returns poor results.
ChatGPT (any model) – converting metadata into .RIS for import into Zotero when no DOI or ISBN is available.
Zoom + Zoom AI transcription – recording supervisory meetings (with consent from all parties).
Image‑generation platforms – predominantly Adobe Firefly and ChatGPT during the study, for discussion with participants.

4  Personal Use

Examples of personal uses (non‑exhaustive):

Claude and ChatGPT (free) – coding as a hobby.
Custom GPTs – e.g. a GPT loaded with all PDF instruction documents for a home solar array and diesel generator, creating a chatbot that quickly answers questions about those processes.
ChatGPT – occasional replacement for search engines. Often more reliable now than Google, since Google search is basically broken.
General experiments – trying interesting or novel generative‑AI platforms, including open‑source options.

5  Non‑Use of AI

Ways I do not use AI, whether by personal choice, ethical concern, or other impediment:

Since 2024 I rarely use AI image generation for blog posts and articles (though I did so extensively in 2023 – 2024). Image generation is far more environmentally costly than text generation, and open‑source images such as those from betterimagesofai.org are usually more than sufficient.
I do not use generative AI with data from my study. At ethics‑application time, too many unknowns made it impossible to claim the technology would be private, secure, and ethical. Business and enterprise models now offer better safeguards, but in my experience large language models code generically, miss nuance, and cannot connect with data on a human level.
I do not use generative AI to write my blog posts; I enjoy writing them myself. While AI might suggest structures for my books and e‑books, those suggestions are usually adjusted or discarded.
I do not use Microsoft Copilot inside Word, PowerPoint, or Excel because it is generally hopeless, even at basic tasks such as formatting.
I do not use Google’s AI Overview, and I will not use Google’s forthcoming AI Mode, which I see as a step backwards for search and online‑content accessibility.
I do not use Grok.
When teaching, I never require students to use AI if they choose not to.

Final Note

While not exhaustive, this disclaimer states the ways I regularly use, or consciously avoid, artificial intelligence and lists the platforms I pay for. I receive no payment from any of these platforms, do not work for technology companies, and am ambivalent about which platforms participants in my professional‑development or advisory contracts choose to use.

With very few exceptions, I do not speak at conferences organised exclusively by a single technology company (though I will speak at conferences where multiple companies provide sponsorships, and those sponsorships help pay speaker fees).

I have no affiliations with any technology or ed‑tech companies. Any platform recommendations I give, including paid ones, are based solely on my own experience; I do not profit from them.