By Dr. Pavan Soni
Not many are aware that both Artificial Intelligence and Design Thinking could be traced back to the pioneering work of the Nobel Laureate Herbert Simon. In 1956, Simon, along with Allan Newell and Cliff Shaw, developed Logic Theorist, the first AI computer program. In his 1969 book, The Science of the Artificial, Simon formally introduced the concept of AI and proposed how problems are solved in the real world, and the role of heuristics (rules of thumb). In the same treatise, he laid the foundation of Design Thinking, stating: ‘Everyone designs who devices course of action aimed at changing existing situations into preferred ones.’
Fast forward to the 21st century. Thanks to the strides made in computation, communication and commerce, both design thinking and artificial intelligence have almost reached the mainstream society and industry. While design thinking is rooted in empathy, AI is often deemed as ‘apathetic’—taking away jobs, displacing industries and wiping-off chunks of the value chains. A pertinent question is – How do you reconcile the two? If you look at the process model of design thinking, you will identify several avenues where AI can lend more than a helping hand. It can, potentially, make design thinking more pervasive and effective without robbing it of its essence of human centricity. Let’s delve deeper into the avenues of convergence.
At a very high level, design thinking could be understood as a human centric, iterative model of problem solving. It comprises five stages: inspire (setting the ‘why’ of the problem), empathize and define (understanding the problem from stakeholder perspectives), ideate (generating high number of relevant ideas), prototyping and testing (validating those ideas in real-world context), and scale (manifesting ideas into products and profits). The entire process is iterative and warrants the following conditions: audacity of goals, ambiguity of context, access to customers, availability of time, and diversity of teams.
Where AI Steps In: Three Key Phases
There are three realms in which AI can aid in the design thinking led problem solving process. Firstly, it’s in the empathize and define stage. Traditional approaches of product design and experience creation relied heavily on human-to-human interactions, ethnography, anthropology, insight clinics, interviews, un-focused groups, and observations, among others, which are costly, time consuming and have a limited reach. Ergo, expansive insights are drawn from limited qualitative samples, hoping that the ensuing ideas would work at scale. Scores of failed products, especially those conceptualized in the west, or drawn from dipstick with the elite, offer testimony to how self-limiting often are the archaic means of design thinking.
With its reliance on data and generative capabilities, AI can help reach a larger audience and augment human intelligence with machine intelligence. Questions can be curated, responses can be analysed in a fine-grained manner, and a more coherent picture can be secured by adopting AI in this fuzzy front-end of problem solving. For instance, Alphabet and Meta regularly scan numerous customer forums and mediums, listening not only to complains but also to what works and what doesn’t, to glean novel insights.
Secondly, in the ideation stage, AI can be of immense value. One of the ethos of creativity is: quantity leads to quality. ‘The way to get good ideas is to get lots of ideas, and throw the bad ones away’, quipped the two-time Nobel Prize winner Linus Pauling. In a typical ideation workshop, participants often get saturated and end-up generating predictable, trite ideas. The situation is further exacerbated with their unwillingness to throw away the bad ones. But what if such ideas are generated by machines? Humans give prompts and machines generate multiple combinations of otherwise commonplace ideas, and together they select the more suitable ones.
Such an approach will not only solve the problem of scarcity of ideas, but also take away the moral hazard of not being able to ‘kill your darlings’. In the realm of new drug discovery, it happens all the time. The algorithms throw at seasoned scientists all possible combinations, which they then sift through to get to the most viable ones. Such machine interventions can lower the Type-2 (false negative) errors, whereby lowering the odds of missing promising ideas that a human mind could not conjure up on its own.
Lastly, AI can be more than time and cost saving during the prototyping and validating phase. Often owing to paucity of time or human biases, not all ideas are put to the test. Consequently, the dominant voices on the table take over, or firms recoil to the tried and tested, low-risk ideas. It tantamount to the wastage of the insight generation and ideation stages, and a dampening of employee morale. What if a lot more ideas could be tested systematically, if possible, parallelly? A lot more data-driven judgement could be taken on their fate. AI can just do that for you, and at compelling budgets. Thanks to Large Language Models and self-learning capacities, the algorithms can identify the right target audience to gauge the efficacy of an idea, ascertain criticality of the sample, and solicit responses. Think of how good machines are in conducting A/B test for a new website launch or placement of your ad on an Insta page.
Let Machines Assist, But Let Humans Lead
With machines willing to aid us, it’s critical that humans move up the value chain instead of clamouring for the same space. The role of Artificial Intelligence in Desing Thinking makes for a compelling case, of which Herbert Simon was rather clear. I hope you give AI a chance in your design journey, without losing the essence of either.
The writer is the bestselling author of the books Design Your Thinking and Design Your Career.
Disclaimer: Views expressed are personal and do not reflect the official position or policy of FinancialExpress.com. Reproducing this content without permission is prohibited.
