All the things AI can do
Will Taylor at UX Brighton 2024
About this talk
There’s an old quote: “There are decades where nothing happens; and there are weeks where decades happen.” And with AI, it can feel like every week brings a “decade’s worth” of new technologies and ideas.
As a UX researcher or designer, it’s hard to keep track of the AI technologies that we can add to our products. So this talk is a primer. It’s a visual, guided-tour of what’s available - what’s “on the menu” - to help you design amazing new experiences for users.
This talk aims to inspire you, help you discover new technologies, and steer your design and research with a clear understanding of what is practical and buildable. To future-proof yourself, you’ll also learn a practical system for staying on top of emerging AI technologies in your field and matching these technologies to your users’ needs.
About Will
Will Taylor is a seasoned designer and engineer who has been designing and building startups for the last 12 years.
He is the founder of Workflow.design, an AI company for the creative industry. Workflow was voted among the top 1% of European startups by Focal ‘24 and has been nominated for multiple design and product awards from Awwwards.
Previously, he was the founder and principal designer at the startup Rota, which was acquired in 2022. Will holds a master’s degree in Theoretical Physics from the University of Oxford and, between startups, is a qualified mental health coach.
Related reading
Books, research, ideas and people referenced in this talk.
Ad · Book titles link to Amazon. As an Amazon Associate I earn from qualifying purchases.
Papers and articles
- Computing Machinery and Intelligence02:32Alan Turing's foundational paper frames the opening question about machine intelligence.
- ELIZA06:43Joseph Weizenbaum's original paper explains ELIZA's pattern-matching conversation system.
- Intelligence without Representation14:37Rodney Brooks describes Nouvelle AI's nature-inspired, adaptive approach.
- Globally Trained Handwritten Word Recognizer15:35A 1993 paper providing technical context for the handwriting-recognition network shown.
- Generative Adversarial Nets21:20The original GAN paper explains one technique behind the synthetic media examples.
- Denoising Diffusion Probabilistic Models21:27A foundational modern diffusion paper for the second generative approach in the talk.
- Attention Is All You Need26:38The original Transformer paper underpins the capability to create and apply knowledge.
Concepts
- Decision trees03:41Explainable branching logic for decisions with human-set rules and outcomes.
- ELIZA effect08:22The tendency to attribute understanding and empathy to conversational software.
- Deep learning19:33Neural networks combined with scale and large datasets to learn from complex inputs.
- Convolutional neural networks20:06Image-classification networks used in the cancer-detection and object-identification examples.
- Recurrent neural networks20:43Sequence models where previous inputs affect later predictions.
People
- Yann LeCun15:35A pioneer of convolutional networks and handwritten-character recognition.
