If Synthetic Users wasn't trying to pitch itself as directly replacing structured research and actual human conversation, it might be on to something. I watched the Synthetic Users talk at UX Brighton last week with some interest. Part of the talk was about the idea of continuous research throughout a product lifecycle. IMO actual human contact has a lot to play in this (eg. usability testing and hypothesis-based design with real people). However, so might computer modelling. The analogy I'm thinking of is how an urban developer might be able to model traffic patterns based on changes to a road junction or traffic light phasing. When provided with enough data about users and an interface - in particular current usage and proposed changes to an interview - a model might help predict unforeseen problems. Not by parroting made-up responses of a few pseudo-interviewees, but modelling responses in aggregate. I feel like that all might be some way off though, particularly with something like an LLM where there is no actual intelligence or understanding involved. To accurately model more marginalised parts of society, surely those members of society would already need to be present to some degree in the training data. And practicalities aside, that in itself might throw up all sorts of ethical questions. That all said, I am drawn to the possibility of modelling future behaviour and reactions, but not as a replacement for real conversations. Benjamin ParryBenjamin Parry (He/Him) • 2nd (He/Him) • 2nd Design Researcher, previously at Clearleft Ltd.Design Researcher, previously at Clearleft Ltd. 3d • 3 days ago I’m trying to find a balanced viewpoint on the topic of LLMs, machine learning, and generative AI in user research. Full disclosure: I’m currently AI sceptical. My concerns are mostly from an environmental and ethical perspective. Finding pro-AI content isn’t a problem. Instead, I’m looking for content taking a critical viewpoint and genuine risks of using AI in the user research practice. I’ve included some examples of content that I consider meets the brief. - Insta-Personas & Synthetic Users by Indi Young – https://lnkd.in/e8syEphM - I tried out SyntheticUsers, so you don't have to by Nilou Salehi – https://lnkd.in/eUtVHMSk Help me understand the risks or convince me I’m wrong. hashtag#AI hashtag#UserResearch hashtag#UXResearch https://www.linkedin.com/in/richardrutter/Richard Rutter