New Gartner survey: Only half of AI models are going into production

Automation and artificial intelligence (AI) are widely embraced by organizations, even though multiple challenges remain, although the challenges may not be what many think.

Across multiple aspects of IT and AI, a lack of qualified IT professionals is often cited as a barrier to adoption. According to a new survey released today by Gartner, a lack of AI talent really isn’t an issue. A whopping 72% of organizations surveyed claimed they can find or already have the AI ​​talent they need.

Everyone builds AI models, but production is harder

While lack of talent isn’t an issue, moving from pilot to production certainly is. Gartner’s research identified a persistent gap between the number of AI models developed by organizations and the actual number being put into production.

The survey found that on average, only 54% of AI models go from pilot to production. That figure is just nominally higher than the oft-cited 53% Gartner reported in a Survey 2020.

“The biggest surprise was the sheer number of organizations that reported implementing thousands of AI models, coupled with the fact that only 54% hit production, and many [indicating] they are not aligned with business value,” Frances Karamouzis, distinguished VP analyst at Gartner, told VentureBeat.

So what does it take to move the needle to move more AI projects from pilot to production? Karamouzis said the one-word answer is discipline.

She says organizations need to take a disciplined approach to aligning with value, ensuring the right talent is in place, and ensuring critical areas of AI trust and security are properly implemented.

Management remains a challenge

The Gartner survey also found that 40% of organizations have deployed thousands of AI models, and that volume creates complexity for governance, as well as tracking the value and return on investment of AI.

The challenge of AI’s lack of governance has been identified in other studies released in 2022. A global research project conducted by Juniper Networks and Wakefield Research, published on June 15, found a lack of maturity in AI governance policies as a barrier to further adoption. However, the Wakefield Research report also found that a lack of talent was a problem, which is not what Gartner sees.

An April 2022 O’Reilly Media report also found governance to be a challenge to AI adoption, with 51% of organizations missing some form of governance plan for AI projects.

The intersection of security, privacy and AI

Security was also not identified as one of the top barriers to adoption by respondents to the Gartner survey. Only 3% of respondents cited security as one of the top barriers, with the top ones being the ability to measure value, a lack of understanding of the benefits and uses of AI, and data accessibility challenges.

But while security hasn’t cracked the top barriers list, AI-related security and privacy issues are rampant, with 41% of organizations admitting they’ve had a problem at some point.

When we delved deeper into the issue of AI security, half of organizations (50%) were concerned about competitors or even partners as risks. However, the real source of risk appears to be insiders. Of the organizations that admitted to having an AI-related privacy or security problem, 60% were attributed to insiders.

“Organizations’ concerns about AI security are often misplaced, as most AI breaches are caused by insiders,” Erick Brethenoux, distinguished VP analyst at Gartner, wrote in a release. “While detecting and preventing attacks is important, AI security efforts should also focus on minimizing human risk.”

The Gartner survey was conducted in October through December 2021, in the US, Germany and the UK. There were 699 respondents to the survey, employed by organizations that plan to deploy within the next three years, or have already implemented AI.