Research shows the humans aren’t totally on the sidelines just yet in AI’s growth.
Our FRESH DATA posts highlight the latest independent research.
One in four enterprises with AI in place has had to rethink, redesign or override an entire AI-based system due to questionable or unsatisfactory results.
This is one of the findings of a recent survey of 307 executives, released by SAS, Accenture Applied Intelligence, Intel and Forbes Insights. (I was part of the team that designed and analyzed the study, as part of my work with Forbes Insights.) While 24% of executives report that such an intervention was necessary over the past year,a majority are still happy with the results of their AI efforts and intend to keep moving forward.
More than half, 51%, say the impact of deployment of AI-based technologies on their operations has been“successful” or “highly successful.” Benefits seen include more accurate forecasting and decision-making (60%), improved customer acquisition (52%), and greater organizational productivity (48%). Two-thirds, 66%, agree that “AI will enable us to mine massive volumes of data faster to inform business decisions.”
For purposes of the survey and report, AI is defined as “the science of training systems to emulate human tasks through learning and automation.”
See also: CIOs often mystified by legacy software, says study
One of the challenges with AI surfaced by the survey is there are still too few audits or instances of monitoring AI-delivered results, the study also shows. The accuracy and relevance of AI can’t be taken for granted, experts warn. “If AI is directly influencing your customers or automating critical operational decisions, you can’t take it for granted that the data will be just right, the models will just work and resultant outcomes will meet expectations,” says Kimberly Nevala, director of business strategies for SAS. “As organizations advance down the AI path and see these very real impacts – both positive and negative – they are moving quickly to put that next level of oversight in place.”
One in five, 20%, report they barely ever review their AI decisions — once or twice a year, tops. At the opposite end, 23% review their AI systems’ decisions at least once a day, if not hour by hour. In the middle are organizations that may conduct reviews n at least a monthly basis.
The question is: who should be monitoring and overseeing AI results? There is evidence that the swift progress of AI is pushing oversight further up the agenda for many organizations — C-level executives such as the chief digital officer, chief data officer or chief analytics officer top the list of answers.
Close to three-fourths of executives, 74%, recognize that close oversight of AI is essential, the survey shows. But there’s still a long way to go before oversight processes catch up with advances in AI. The bottom line is that while AI holds a lot of promise — and is the hype topic of the moment — enterprises need to avoid the temptation to dive in head first.
