Survey: AI Must Be a ‘Strategic Lever’
Hyper-scalers are racing to deploy AI technologies. Meanwhile, other sectors such as finance, healthcare and transportation ramping up AI initiatives are confronting challenges while considering precisely where automation technologies fit into individual company strategies.
That along with lingering concerns about AI hype are among the key conclusions of the latest enterprise AI survey, this one released this week by business consultant KPMG. Along with getting a handle on AI capabilities, the survey of about 750 senior executives identified lack of training and initial investment shortfalls as the biggest AI deployment challenges.
“Executives need to look at AI as a strategic enterprise-wide initiative, not simply a technology play,” said Traci Gusher, KPMG’s principal AI analyst. “It’s not just about installing AI technologies,” Gusher added. “It’s about using AI as a strategic lever to transform the business. And that requires building deep AI capabilities across the organization—both from the bottom up and the top down.”
The survey also uncovered a paradox: The technology and transportation sectors that have been at the forefront of AI deployments—machine learning for hyper-scalers and autonomous systems for car makers—expressed the most skepticism about machine intelligence. Sixty-nine percent of transportation managers said AI has yet to move beyond the “hype” phase while 57 percent of tech executives expressed doubts about AI’s potential.
Financial services executives were the most bullish industry sector when it comes to AI deployments, citing applications like fraud detection and the eventual automation of traditional banking services.
The survey also surfaced familiar issues such as the need to develop enterprise-wide AI tools. “AI can’t be built in a silo,” KPMG’s Gusher stressed.
That approach is seen as one way to faster return on investment. Those investments extend beyond R&D to include training, attracting scarce AI talent and tackling data security and privacy issues.
Training also goes hand-in-hand with concerns about job security, the survey notes. As more business processes are automated, workers will have to be retrained to handle high-level task such as data analytics.
The next critical step for analysts and decision-makers will be establishing trust in the algorithms churned out by their models. Hence, the AI survey stresses the importance of explainable AI. “We must ensure the integrity, fairness, explainability and resilience of our AI models,” Gusher said.
The survey also highlights the need in at least some sectors for greater regulation. Car makers in particular said regulators must help establish AI standards and safeguards, with 82 percent endorsing a government role. Meanwhile, 77 percent of transportation executives said AI represents a “direct threat” to consumer data security and privacy.