In recent years, we have focused increasingly on diversity and inclusivity: How do we ensure that all colleagues and customers feel treated appropriately? How do we write our documents? What do we look for in our software and UX design? What influence does our language have?
Now the use of large language models has grown rapidly, and we feel that we can't do without AI. At the same time, we know that these models contain many biases that can lead to ethically questionable or, in human judgement, unfair results - and that AI has learned these biases from us! But if AI is learning from us and we are learning from AI, where do we start to improve together?
In this talk we will look at some examples of inclusive design and diversity in software development. Building on this, I will show how I incorporated these examples into prompts when building a CustomGPT to detect diversity bias in documents or websites - and what I learned from it.
We used to ask: 'How does our thinking change our language and our language change our thinking?'
Now we need to ask: 'How does our language change AI and AI change our thinking?'