AI, governance

If it’s Friday, it must be in Austin, TX.

“Every business used to want to be a tech company. Now, it’s AI,” says Michael Housman, a data scientist who in fact is home in Austin on Friday but is packing his bags for London and Dubai. He’s already made pit stops in Asia (Singapore and India), Europe (Italy) and North America (Chicago and San Francisco Bay Area.)

Forgive Michael Housman if he forgets which continent he’s on. A 10-week tour of the world, discussing the state of generative-AI with IT buyers, business and thought leaders, and educational institutions, can do that to people.

Housman is globe-trotting at the behest of corporations, business professionals and educators to help “demystify and unpack AI.”

“Business leaders, lawyers, accountants, tax preparers are nervous about their jobs. They want to know how to use an LLM, and how is AI going to change this business and not get left behind,” he said. “How do these tools replace or augment my workers? How does this disrupt my space? The internet killed Blockbuster video, social media hurt newspapers. The fear is palpable.”

The most frequently-asked question Housman gets is, he said, “Will I lose my job?”

“Well, it will evolve and change the nature of your job,” Housman said. “New technology creates jobs in adjacent areas, dating to the steam engine, printing press and automated switchboard. Word processors replaced proof readers. The same will happen with AI.”

Bottom line, he says AI adoption is in the “very early stages” and there is no guarantee for overnight success.

Case in point: A recent report from market researchers Genpact and HFS Research found only 5% of enterprises have mature gen-AI initiatives, yet 61% of executives allocate up to 10% of their tech budgets to hasten adoption.

Then, there is the chasm between AI tools and worker productivity.

The disconnect plays within a narrative that, in rushing technology to users, companies assume their job performance will immediately skyrocket once they adopt the newest, shiniest object. Initial results, however, prove otherwise with AI.

An enduring myth persists, as with previous technologies, that AI adoption will be easy, leading to optimum results. Despite overwhelming confidence from 80% of users that AI will super-charge their productivity, only 9% of American workers say they are getting the results they need out of AI on their first try, according to a new study of more than 1,000 workers nationwide by digital-adoption platform WalkMe. [About 41% of those polled claim they use AI tools daily.]

“It is an interesting concept of false confidence among workers in everyday use of AI. They’re confident they are getting the best results when it isn’t the case,” said Rob Baran, group manager, product marketing, at WalkMe. “This happens with most technologies — as was the case with enterprise SaaS over the years.”

One reason may be that the study found only 7% of U.S. workers said they got extensive AI training in their employer’s haste to put AI to use.

It is only on repeated attempts that so-called resilient users, usually through trial and error, get better results, said Baran. And those who don’t get results, quit. [The study also discovered that 86% of respondents are more likely to use AI tools if the tools proactively guide them in the flow of work.]

Of course, more than 80% of data leaders lack a plan for gen-AI use cases, based on a survey of 700 of them at the recent Snowflake Summit.