AI vs Humans? An experiment.
Can AI beat humans when it comes to Innovation and Human-Centred solutions?
Last week DesignThinkers assembled an amazing diverse group of Africans from a variety companies and countries learning about Design Thinking from expert facilitators, coaches and “Dragons” while working on a challenge related to the UN’s 6th Sustainable Development Goal - Ensure availability and sustainable management of water and sanitation for all.
This “Learning by Doing” event had a secret challenge running in the background. Can artificial intelligence (AI) design human-centred and innovative solutions equal to or better than the participants of the DesignThinkers 4-day Bootcamp?
Here is how it was orchestrated:
· Two multi-disciplinary teams followed a typical Design Thinking process to explore the problem space and design potential solutions (their focus was on the desirable component, not having much time to test feasibility and viability).
· At the same time (although it took much less time!), one of our colleagues Jennifer Sutherland, took the same challenge as the participants and plugged specific questions into various AI tools in order to simulate the Design Thinking process.
The challenge statement: was “How might we ensure the continuous supply of quality water to households (and businesses) in and around Cape Town?
Read on below to find out more…
How it transpired
The 4-day DesignThinkers African Bootcamp incorporates a “learning by doing” approach to enable participants to get a deep sense of how “pressure cooker” innovation feels when working in diverse, multi-disciplinary teams.
Participants were guided by their coaches to focus on Design Thinking mindsets and principles to enable them to research and emerge a fresh and deeper understanding of the challenge.
Forming, storming, norming and performing
The first day was intense as it involved getting to know new people and interpret the broad challenge provided by DesignThinkers Academy. The participants were teamed with people they didn’t know and therefore needed to spend time learning each other’s innovation style.
The challenge the participants received was deliberately broad to enable the team to diverge and discover new opportunities and unmet needs for a target group not yet explored.
Participants explored the unknown and learnt how to engage stakeholders with an open mind and curious attitude while dropping biases and assumptions to discover the pain points for water users.
Diverging: Empathising and Sensemaking
Day 2 is the empathy and research day. It involved finding nuggets (insights around the latent needs people have). And then making sense of the complexity related to people’s daily lives, their challenges and pains in and amongst the water-energy nexus. This day was the most intense as people went out and spoke to strangers, trying to understand what people do and why.
Jani de Kock our Senior Facilitator led the teams, helping to bring their minds to interpret and understand behaviours in relation to people’s goals and motivations.
Converging: Reframing and Ideating
Day 3 involved refining the personas, empathising further using Journey Mapping and emerging the main insights and opportunities. The participants connected to their persona’s preferred outcomes and then built a reframed problem statement focusing on the underlying emotional and functional needs.
From here a set of ideas were elaborated on, resulting in a multitude of concepts to choose from. Participants were encouraged to focus on small wins and immediate needs with a long term view in mind.
Prototyping and pitching
The temperature in the “kitchen” went up on the final day when the participants were asked to prototype and prepare for their pitch. The Dragons were on their way to assess the concepts.
Teamwork and alignment became crucial and final concepts were settled on, with both teams converging on slightly different personas, but both using an App to support the city in accumulating data and educating and informing their personas on their behavior and consumption patterns.
The Dragons found that both teams could have connected closer to pressing human needs of the personas they developed, and some new iterations of their solutions were suggested.
How did the AI experiment fare?
The AI tools (primarily ChatGPT) were given very similar prompts to the human participants. The results were surprisingly deep at times while being rather robotic-sounding at other times. The AI tool selecting a low LSM persona but the AI itself admitted to not being able to gain sufficient empathy through engagement with users and stakeholders. The ideas generated were relatively diverse so it was disappointing that the AI also “voted” to go with the app. It was interesting that the AI generally kept track of the conversation and didn’t need concepts like “dot voting methodology” explained.
Time constraints meant we couldn’t explore prototyping AI tools so the AI generated concept was shared in PowerPoint format. It was closely aligned to the findings of the two separate teams, however, did not adequately address the emotional components
The DesignThinkers African Bootcamp is a baptism of fire, putting the participants into a pressure cooker with team dynamics and self-awareness (EQ) at the forefront while working in an interdisciplinary fashion.
In conclusion, the iterative design process needed to complete a truly human-centred solution where shared value is created needs more time than the 4 days of learning a framework and applying it at the same time provides. Secondly, AI, while able to converge to similar solutions as humans, failed (unsurprisingly) to show the same level of empathy and customised design that was displayed by the two teams. As with user research, to use AI effectively one needs to focus heavily on asking the right questions. Having said that, there are certainly times where AI could add value to Design Thinking practitioners – like quickly constructing a draft research plan and brainstorming ideas.