Teaching AI thinking and design thinking for the age of AI

Executive overview

AI is already displacing jobs and will accelerate. The question is not whether to engage with AI, but how to stay relevant alongside it.

Human advantage lies in zero-to-one innovation — creating what doesn't yet exist. AI excels at optimisation within known problem spaces; humans must focus on defining new ones.

The core survival strategy: develop AI thinking to know where humans add value, then use design thinking to systematically generate that value.

The three components of AI thinking

• Understand how AI works: rule-based systems gave way to data-driven models like deep learning and reinforcement learning
• More data generally produces better AI solutions
• Distinguish human ability from machine ability — know which tasks to delegate
• AlphaFold (DeepMind) predicted structures for nearly all known proteins in two years; humans had found less than 1% in decades
• Once machines can do a job, let them; redirect human effort to what machines cannot do
• Use AI as a tool to accomplish higher-order work, not a threat to avoid

Design thinking: a systematic path to innovation

• Creativity can be taught through a repeatable methodology
• Design thinking does not make you Steve Jobs — it makes you more creative than your current self
• Step 1 — Empathise: understand who you are inventing for and their emotional experience
• Step 2 — Define: identify the right problem; most failures start with the wrong problem
• Step 3 — Ideate: brainstorm broadly, gather user feedback, select ideas worth prototyping
• Step 4 — Prototype: build a version to test with real users
• Step 5 — Test and iterate: all design is redesign; expect to loop back to earlier steps

The Nepal incubator case

• Original brief: get more incubators into Nepal
• Team assumed the problem was equipment shortage — designed solutions in California
• On the ground, they found expensive incubators already in clinics, unused because staff didn't know how to operate them
• Real problem: families in remote villages needed cheap, home-use incubators
• Redefining the problem led to a successful, low-cost solution
• Lesson: empathy and field research must precede definition, not follow it

Responding to ChatGPT in education

• ChatGPT surprised people by conversing naturally and writing code — AI is smarter than expected in many domains
• Blocking AI tools is futile; technology adoption cannot be stopped
• Assigned students to write an essay using ChatGPT, then discussed how they used it and what they noticed
• Goal: understand AI's impact on learning, not avoid it
• Teach students to innovate with AI, not compete against it