Manufacturers slow Gen AI rollout due to accuracy concerns

According to a study by Lucidworks, manufacturing companies are slower than planned in implementing generative initiatives in the field of artificial intelligence. The study surveyed over 2,500 decision makers globally who are in a position to make technology-related decisions and found that 58% of manufacturing executives plan to increase their spending on artificial intelligence in 2024.

This percentage is lower than the global consensus (63%) and the US consensus (69%).

The percentage has also fallen significantly. In 2023, 93% of all executives and 93% of manufacturing executives planned to increase their spending on AI.

Conversely, the Lucidworks study found that nearly 50% of manufacturers worldwide reported an increase in cost savings after implementing AI initiatives this year.

No room for error

A 36% of all respondents expressed concern about the accuracy of answers due to hallucinations, but in the manufacturing sector this figure rose to 44%.

Manufacturers cannot afford to implement AI systems that can cause costly errors, such as incorrect pricing, recommending competitor products or selecting the wrong components.

Although only 20% of planned AI projects were implemented by manufacturers last year, 55% believe they are on par with their competitors in implementing AI.

"While many manufacturers see the potential benefits of generative AI, challenges such as response accuracy and cost are causing them to take a more cautious approach. This is also reflected in their spending plans. Significantly fewer manufacturers are planning to increase their investment in AI than last year," said Mike Sinoway, CEO of Lucidworks.

"However, the above-average cost savings in 2024 could make them more optimistic next year. B2B companies and manufacturers have a lot to gain by finding a balance between cost and risk to increase efficiency, improve customer experience and reduce operational costs through generative AI," he added.

In 2023, 70% of manufacturing companies prefer more expensive commercial AI models. The company believes there could be a shift towards more accessible open source models if they prove to be more efficient and innovative at a lower cost.