Why can't our stroke leaders figure out a way to do this? Maybe because we have NO stroke leaders that can think their way out of a paper bag. Obviously our fucking failures of stroke associations need a lot of help solving stroke, why don't they acknowledge how fucking awful they are at stroke?
Crowdsourcing Our Cognitive Surplus
Tapping into the intelligence of groups—within a company or around the globe—can help organizations combat bias, make better decisions, and compete for talent and ideas with the help of artificial intelligence.
Think of crowdsourcing as applying the principles of the sharing economy to cognitive surpluses. Many people have thoughts, ideas, and skills with real business value that often go unused. Companies can tap into those surpluses both internally and externally, often with the help of technology.
Better Decisions
Fifteen years after the publication of Michael Lewis’s “Moneyball,” hiring remains one of the biggest areas where economically important decisions are subject to unreliable instincts. Many HR professionals quickly sort through resumes, selecting the ones that seem promising based on simple rules of thumb and gut feel. Typical job interviews are heavily influenced by first impressions—and notoriously unreliable indicators of future performance.
Harnessing the collective intelligence of multiple interviewers is a promising path forward. In the book “Work Rules!” Google’s former head of people operations, Laszlo Bock, writes, “None of us are as good at interviewing as we think we are. On our own, most of us make hiring decisions that are less reliable than the wisdom of the crowd. We’re also susceptible to confirmation bias. We’ll look for signs that confirm our unconscious beliefs about the candidate.” Technology can facilitate this sort of collective intelligence. For example, the “Applied” software platform, developed by the behavioral scientist Kate Glazebrook, combines the assessments of multiple interviewers, operating in an environment designed to mitigate cognitive biases, to improve hiring decisions. This can lead to lower turnover, improved performance, more diverse workforces, and better business outcomes.
Better Ideas
Crowdsourcing can enable not only better decision-making but better ideation. When trying to decide on a new business strategy or product, the best idea may come from a serendipitous encounter, an observation from a front-line worker, or a cutting-edge concept from a junior team member fresh from university. Efficiently eliciting such input can be challenging, however.
Wiki surveys—developed by the pioneering computational social scientist Matthew Salganik—illustrate how technology can enable better crowdsourcing of ideas. In New York, the Bloomberg administration used a wiki survey to generate new ideas for greening the city. The survey was initially seeded with ideas from a committee of experts, and then opened to the public. Respondents voted on the relative merits of the various ideas. If survey takers had ideas that were not already included in the current list, they could add them for consideration by subsequent survey takers—the wiki aspect of the survey. Only three of the 10 ideas on the final prioritized list came from the original survey; the other seven—including the five highest scoring ideas—were contributed by ordinary citizens. One might call this “democratized ideation.” Note that Salganik’s wiki survey software is open source, so other organizations can easily try it.
Better Creativity
Top advertising agency Tongal has already figured this out. It employs a fraction of the full-time employees of a traditional agency and charges its clients much less. It achieves this feat by relying on more than 125,000 people who generate ideas, storyboards, and ads for brand and video content via a crowdsourced model.
Crowdsourcing can be not only a more cost-effective approach to deploying creative workers—it may offer the opportunity to elicit ideas and foster co-creation with a company’s customers. When Lego wanted to create a series of stop-motion animation videos, it launched a crowdsourcing competition. The winning video came from a 14-year-old target consumer, Zach Boivin. “His work is nothing short of spectacular,” says Caleb Light-Wills, senior vice president of product at Tongal, through which Boivin operates. “The depth he achieves in three-minute videos is amazing.” Boivin has since gone on to work on many more projects for the company.
Crowdsourcing can enable companies to locate passionate and talented individual contributors they never would have found through their traditional hiring processes.
Better AI
Hiring and contracting are the two classic ways companies gather human capital to solve problems. A third model may increasingly be deployed—parallel processing hundreds of skilled resources via crowdsourcing—and Google may be showing the way.
In early 2016, Google CEO Sundar Pichai wrote an open letter asserting that AI research projects were an area in which Google could set itself apart from competitors and reaffirming the company’s commitment to AI investment. For Google and others, however, an aggressive AI strategy is contingent on scalable access to scarce data science talent.
Less than a year later, Google announced its decision to acquire Kaggle. The company, founded in 2010, is an online crowdsourcing community with data scientists and specialists spanning 194 countries, including some leading experts in the field. Kaggle hosts challenges in which crowds of data scientists, engineers, mathematicians, and other specialists compete to develop solutions that meet or exceed defined performance criteria set by their clients. In addition, community members can collectively investigate insights from open data sets to improve their data science chops. Intel now includes participating and winning in Kaggle challenges as a qualification for applying for data science jobs.
Google’s acquisition will combine a powerful machine learning cloud with one of the world’s largest data science crowdsourcing companies. Prior to the acquisition, Google started working with Kaggle on a $100,000 machine learning challenge to classify YouTube videos; as of now, 334 data science teams are competing for the prize money. Kaggle is expected to move under Google Cloud but maintain its brand and continue to run open competitions.
According to Experfy Founder and Co-CEO Harpreet Singh, “As AI begins to show its disruptive potential, AI’s mainstream adoption is giving rise to a severe scarcity of data science talent. What we are seeing is that the best data scientists often eschew 9-to-5 jobs. Instead, they are freelancing and can be found all over the globe. What Google probably realizes is that such platforms as Kaggle provide one of the most efficient means to access highly specialized AI talent.”
*****
It’s still early days for crowdsourcing. Companies are just beginning to get a handle on the kinds of problems collective intelligence can help them solve—and what tasks are not suited to the approach. In an era when so much critical talent—from creatives to data scientists, developers, and cybersecurity personnel—is hard to find, crowdsourcing offers a valuable and often cost-effective option. It can help companies fast-track important work, generate innovative ideas, and harness the intelligence of talented individuals around the world.—by Jim Guszcza, chief data scientist, Deloitte Consulting LLP; and Balaji Bondili, senior manager, Deloitte Pixel, Deloitte Consulting LLP
No comments:
Post a Comment