Many companies expect in the near future to derive substantial benefits from adopting artificial intelligence-powered technologies. However, they must first overcome some significant obstacles.

In a global survey of 300 C-level executives and their direct reports by consulting firm Protiviti, 16% of the respondents say their companies are getting “high” or “very high” value from their investments in AI. But more than three times that many (52%) say they anticipate they will be doing so in two years.

Protiviti identified five major barriers to AI adoption.

First, although improved cybersecurity is a major advantage of AI, it also brings its own cybersecurity risks due to the greater access to sensitive and personal data.

The Future of Finance Has Arrived

The pace with which finance functions are employing automation and advanced technologies is quickening. Rapidly. A new survey of senior finance executives by Grant Thornton and CFO Research revealed that, for just about every key finance discipline, the use of advanced technologies has increased dramatically in the past 12 months.

Read More

“When you have regulations such as the recent [GDPR] in the European Union, the stakes get higher,” says Protiviti managing director Tryone Canaday. “You need to know that these algorithms are using personal data in both a legal and ethical way, and that they are thoroughly protecting that data at the same time.”

Second, there are other regulatory constraints. Organizations must be able to audit their advanced AI applications, not only to ensure that data is secure, but that company leaders understand how the applications work.

Deep learning (DL), a higher form of machine learning that applies successive layers of representations to teach computers to understand and make decisions by themselves, is a particular concern when it comes to auditing.

“When you start to look at things like DL, where computers are essentially writing their own algorithms, how do you audit that?” says Canaday. “Can a regulator have confidence even when some people don’t understand how AI is coming up with its answers?”

Third, only half of companies apply as much rigor to AI business cases as they do to other investments. About one in three bases AI investments solely on proof of concept, while one in five requires neither ROI nor proof of concept.

Fourth, however, compelling proofs of concepts and pilots are essential, because senior executives simply remain skeptical about advanced AI. Among survey participants, only 8% of CEOs, 14% of COOs, and 21% of CFOs saw AI as considerably or very important to the future of their business.

Why? One reason is media coverage of AI, which is often negative. Respected voices have amplified the resulting concerns. For example, in 2016 renowned physicist Stephen Hawking famously said, “The rise of powerful AI will either be the best or worst thing to happen to humanity.”

Unsurprisingly, many technology executives disagree with their non-technology counterparts: 57% of chief technology officers and 46% of chief data officers recognize the importance of advanced AI adoption.

Protiviti advises avoiding huge infrastructure investments early on. “Pilots should address low-hanging fruit that is not easily solved by other analytics approaches,” the survey report says. However, choosing the right applications can be difficult as companies deal with legacy systems and other infrastructure issues, it notes.

Fifth, universities are not producing enough advanced AI specialists, which is spawning a talent war and pushing up salaries.

Moreover, Canaday says the ultimate challenge for companies is finding people that understand both the business and the technology and data. Often they are strong in one or the other. “You need people who can connect those circles,” he says.

, , , , , , , , , ,

Leave a Reply

Your email address will not be published. Required fields are marked *