Course Outline
 
This workshop introduces practical steps to responsibly integrate generative AI tools like Copilot into data-driven work, with a strong focus on data ethics and a statistician's understanding of how these tools operate under the hood. The aim is to move beyond simple "prompt engineering" to equip participants with the skills to critically evaluate and validate AI outputs for accuracy and appropriateness.
 
Learning Outcomes
 
Participants will be able to:

• Understand the underlying statistical methodologies of generative AI tools, such as next-token prediction.
• Identify the statistical reasons behind AI hallucinations, bias, and regression to the mean.
• Apply a statistician's critical eye to test the robustness and validate AI-generated text outputs against source data.
• Integrate generative AI tools responsibly and ethically into data analysis and statistical tasks.
• Use AI tools to improve the accessible communication of data and decision-making.

 
Topics Covered
 

• How AI tools decide what to do with your prompts and data: Explore the underlying statistical methodologies behind how post-2023 AI tools deconstruct input and generate outputs, specifically the mechanics of "next-token prediction" as an exercise in conditional probability and statistical distributions.
• The Statistics of Hallucinations and Bias: Investigate the statistical reasons for AI confidently lying, including hallucinations, regression to the mean in AI outputs, and how training data distributions create inherent biases.
• Critiquing AI Outputs like a Statistician: Learn techniques to test the robustness of an AI summary, interrogate its confidence, and validate its outputs against the source text, moving beyond simple prompt engineering.
• Communicating your data and decision-making: Use AI to help communicate data accessibly to drive impact.
• Responsible and ethical use of AI and data science:  how to ensure data is sourced, processed, modelled, and communicated in a responsible and ethical way.

Charlotte Hadley has been working in data science and communication since 2012. Charlotte has not only built data teams and services, she has focused her career on training users and groups in data methods. Charlotte is the author of LinkedIn Learnings most popular data science course in R, she’s been a visiting lecturer in data science and ethics, and built-up Admiral’s multi-award winning Data Academy catalogue and train-the-trainer program.

Fees		Registration before   17 December 2026	Registration on/after 17 December 2026
Non Member  RSS Fellow  RSS CStat/Gradstat/Data Analyst   also MIS & FIS		£535.00 +VAT  £455.00 +VAT £430.00 +VAT	£565.00 +VAT  £480.00 +VAT £450.00 +VAT
Group discounts are also available*: 3-5 people 6-8 people 9+ people *Discount only applies to non-member price		10% discount 15% discount 20% discount