Crowdsourcing is a great way to generate a wealth of new ideas, but filtering out the bad suggestions can often be costly and inefficient.
Artificial Intelligence could, however, soon make crowdsourcing obsolete, according to a new study published in the INFORMS journal Marketing Science. The research, led by a team from the University of Oxford, Universidad de Los Andes in Chile, the University of Southern California, and the University of Innsbruck, Austria, developed a simple AI model that was able to efficiently screen out bad ideas in the crowdsourcing process, leaving only the most promising ones for consideration.
The AI model could also be easily adjusted by managers to determine the number of ideas to be screened out, without losing potentially good ones.
The fuzzy front end
“Idea generation and screening are fundamental to marketing success because they comprise the start of new product development,” said Gerard Tellis, one of the study authors of the peer-reviewed article ‘Can AI Help in Ideation?’ “They belong to the ‘fuzzy front end,’ a key point of leverage in new product development.”
The researchers tested the AI model using data from Hyve, an innovation company that runs a crowdsourcing platform. Using a data set of 21 crowdsourcing contests, which included 4,191 ideas, the model was fitted on 20 contests and used to predict success in the 21st idea left out.
They found the best model could screen out 44% of ideas, losing only 14% of good ones, while a two-step approach could exclude 21% of ideas without losing any winners. A new predictor, word atypicality, proved efficient in screening out atypical ideas, favoring more inclusive and rich ones.
The study found that AI models, once developed, prove to be relatively low-cost to operate and – importantly – do not succumb to internal biases. They are also transparent and offer privacy, helping to safeguard a company’s intellectual property.
That doesn’t mean that AI can handle it all. “People and experts are still needed,” said Johann Füller, another author of the study. “In the selection phase, AI can replace humans in the screening and narrowing of those ideas. But in the long run, if automation is used properly, it can even eliminate the need for human idea generators and make crowdsourcing itself obsolete.”