Science

Organizations don’t trust AI enough to forego human decision making

Fivetran announced survey results showing that while 87% of organizations consider artificial intelligence (AI) vital to the survival of their business, 86% say they would find it difficult to fully trust AI to make all business decisions without human intervention. 90% of respondents say their organizations continue to rely on manual data processes.

Vanson Bourne’s online survey of 550 senior IT and data professionals in the US, UK, Ireland, France, and Germany also found that only 14% of organizations rate their IT maturity level as “advanced”, meaning they use universal applications. AI. automatically make forecasts and trading decisions. 41% of respondents admitted that the way they use AI in their organization could be improved. That number jumped to 64% when considering only US respondents.

“This study reveals significant gaps in the efficient movement and access to data between organizations. A successful AI program depends on a robust database, starting with a cloud storage or data lake as the foundation,” he said. George Frazier, CEO of Fivetran. “Analyst teams using the modern data stack can more easily add value to their data and maximize their investment in AI and data science. »

Inefficient data processes reduce AI breakthroughs and revenue growth

Organizations appear to be laying the groundwork for more complex AI projects and plan to spend 13% of their global annual revenue on this over the next three to five years, up from 8% today. Nearly all organizations surveyed already collect and use data from operating systems, but their ability to use that data for AI models is hampered by deep data issues:

  • 71% struggle to access all the data they need to run AI programs, workloads, and models.
  • At least 73% find every step of extracting, loading and transforming data to the point of turning it into actionable advice for decision makers challenging.

These inefficient data processes force companies to rely on human decision making 71% of the time. Ineffective AI programs are also hurting organizations financially, with respondents estimating that they lose an average of 5% of their global annual revenue due to models built with inaccurate or poor-quality data.

AI talent remains unused

The prevalence of poor-quality, siled, and outdated data means that data scientists working in all of the large organizations surveyed spend less than a third of their time building AI models, with the rest devoted to tasks unrelated to their job. .

As a result, 87% agree that data scientists in their organizations are not using their full potential. However, hiring is cited as the biggest hurdle to AI adoption (39%), highlighting the responsibility of organizations to urgently leverage the talent they already have.

Back to top button

Adblock Detected

Please consider supporting us by disabling your ad blocker.