This conversation with Sentient Design authors Josh Clark & Veronika Kindred really stuck with us.
AI is proving to be extremely powerful at many things. It also has many tasks that it performs poorly at.
One of the strengths that's often under-discussed is AI's power as a tool for divergent thinking. When designers want idea and inspiration we're happy for the tool to make up bullshit. The more outrageous the stimuli, the more we want it to challenge assumptions.
This holds true beyond creative tasks too.
When summarizing and making action items sometimes there's value in challenging the default expectations of the LLM. You can prompt tools to frame outputs from specific perspectives or give a list of unique circumstances to reframe outputs against.
By challenging LLMs to come up with new perspectives they can enable us to avoid a default view. Remember, LLMs have been found to output what they think we want to hear. They're trying to appease us and give us the most acceptable answers.
It's out job at AI operators to attach additional documentation and include vitally important scenario framing into our prompts so we don't just get the most obvious answer.
Force the system to give us a new way of organizing complex project.
Challenge the system to make content that generates intense feelings.
Push its capabilities of analyzing data from the perspective of a micro-segment.
Listen to the episode with Josh & Veronika here: https://podcasts.apple.com/us/podcast/19-authors-of-sentient-design-ai-powered-self-aware/id1734499859?i=1000672422070
Get their book here: https://rosenfeldmedia.com/books/sentient-design/
Related Videos
#servicedesign #systemicdesign #journeymapping are the ideal tools to pinpoint how to leverage #AI
A lot of vendors sell LLM solutions as the great fix that your company needs, but the technology is often under-performing because its the wrong solution to your problem. The deeper we get into the AI...
![Over reliance on AI is a Problem [Brittany Hobbs & Arpy Dragffy] - Design of AI video podcast](https://images.podcastpage.io/fetch/https%3A%2F%2Fi.ytimg.com%2Fvi%2FujMByL9e0wI%2Fmaxresdefault.jpg?w=400)
Over reliance on AI is a Problem [Brittany Hobbs & Arpy Dragffy] - Design of AI video podcast
Employers and the media are pressuring us to adopt AI quickly. But overreliance on large language models poses huge risks. Not only is GenAI highly likely to make mistakes, we’re putting the entire da...
![32. Overreliance on AI is a Problem that only Expertise can Solve [Brittany Hobbs & Arpy Dragffy]](https://images.podcastpage.io/fetch/https%3A%2F%2Fi.ytimg.com%2Fvi%2FnJ95QPMPzrM%2Fmaxresdefault.jpg?w=400)
32. Overreliance on AI is a Problem that only Expertise can Solve [Brittany Hobbs & Arpy Dragffy]
Employers and the media are pressuring us to adopt AI quickly. But overreliance on large language models poses huge risks. Not only is GenAI highly likely to make mistakes, we’re putting the entire da...