As someone who has provided tax advice for years, especially advising accountants and lawyers who do not have their own in-house tax team, the CIOT AI for Tax course was both timely and thought provoking. I approached it not as a tech expert, but as a business owner wanting to understand where the opportunities and risks really lie.
From the first module I appreciated how clearly the course sets expectations. It doesn’t pretend AI is a magic wand. Instead, it lays out the differences between hype and reality. It guides you through what machine learning, generative models, and natural language processing actually do, and what they struggle with. This was so helpful in ensuring we treat AI tools with healthy scepticism rather than blind faith.
One of the biggest focuses for me was around risk and oversight. In our ‘old school’ world, risk is usually legal, reputational or financial. AI adds a new dimension: algorithmic risk. How are models trained? What biases are baked in? Who audits them? The course underscores the importance of having humans in the loop even when automation steps in. As advisers, we need to ask probing questions of AI vendors, demand transparency, and retain ultimate responsibility for outputs.
Another key takeaway was that AI isn’t here to replace the adviser, but to amplify what we do, making it faster and more efficient. The modules covering document processing, intelligent extraction, search, and summarising of data showed tools that can shave hours off manual work. But the secret is in how you integrate them: using AI to free up time for deeper analysis, strategy, client conversations, rather than pushing it into every corner for its own sake.
What I also liked is that the course doesn’t shy from the limitations and issues arising from AI. It reminded us that the human judgment, the ethical safeguards and the professional accountability must still sit at the centre. Even in a world of algorithms, tax advice remains a human service.
Since finishing it, I feel more confident explaining AI’s strengths and limits to others. I’m more inquisitive and knowledgeable when assessing new tools. I’m not starry eyed about AI, but I also don’t fear it. It’s a tool, and how effective it is depends on how wisely we use it.