The single most important factor to successful AI adoption

The key to rebranding as an artificial intelligence (AI) company is tapping into an “ultra-specialized talent pool” with a deep understanding of the field, says the chief science officer at Royal Bank of Canada.

“It’s impossible to overstate the importance of this expertise level,” said Foteini Agrafioti, RBC’s chief science officer and head of Borealis AI, the bank’s research and development arm. “If you want this [AI] capability in your organization, you have to hire the people with the perfect balance of data intuition and state-of-the-art knowledge. These people are almost all academics.”

Machine learning models run on mathematics whose subtlety requires a deep understanding of data domains. Simple tasks can be daunting to those without the right experience. “Mistakes in the machine learning space are extremely common, but when applied to business they can have real-life consequences too – up to and including life-and-death,” wrote Agrafioti in How to Set Up an AI R&D Lab, published Tuesday in a Harvard Business Review blog.

She used a 45-year old “textbook example” of how easily these mistakes can occur. In 1973, the University of California at Berkeley compiled their graduate school admissions figures and discovered what appeared to be a significant bias against women applicants: 44% of male applicants were admitted to graduate programs versus 35% of female applicants.

Fearing a lawsuit, school officials shipped the data to their statistics department for a closer look. When analyzed by parsing individual departments, the data showed women were applying to harder, more competitive programs with lower admission rates than their male counterparts. Actually, women had slightly higher admission rates than men per department.

“Researchers will be familiar with this phenomenon, known as Simpson’s paradox,” Agrafioti wrote. “Machine learning challenges have significantly increased in complexity since then and it takes years of training and experience to develop a well-honed institution that can sniff these problems out.”

In the case of Berkeley, the administration had access to an in-house team of top-notch statisticians. “But it’s easy to see how inexperienced analysis could have tied the university up in a costly, lengthy legal battle. Instead, the university emerged with a more nuanced understanding of its admissions process.”

What business leaders tend to miss, Agrafioti said, is that the scientific progress made in AI does not automatically render the technology ready for any environment. Each business carries its own unique challenges and requirements, from proprietary data types to operational constraints and compliance requirements, which may require additional customization and scientific progress.

The best way to de-risk the pursuit of AI is to pair fundamental research (pure science) and applied research (designed to solve specific real-world problems; what applied researchers bring to the table is knowing when to stop researching and focus on delivering a solution). In RBC’s case, fundamental research aims to advance natural language processing – the field of AI that can understand language – to a place where it can independently perform high-level language-based reasoning and grasp complex relationships at the same level humans do. Applied researchers then ensure these solutions can become immediately applicable to financial services.

The last step is creating the right environment. Providing a dynamic, comfortable and unique workspace is a good start; this includes reproducing some of the working conditions researchers have brought from academia.

“Investing in big machinery won’t cut it,” Agrafioti wrote. “To truly have impact and remain relevant in the market, it’s up to executive leaders to build the bridge between research and commercialization. Only in this collaborative vein will AI’s true impact flourish.”