Research by an Intel subsidiary revealed hype around generative AI (GenAI) was yet to be matched by deployments, with a majority of executives surveyed apparently reticent due to challenges around rolling out large language models (LLMs).

The survey of 430 technology professionals by Intel company cnvrg.io found 10 per cent stated their organisation had launched generative AI (GenAI) set-ups in 2023.

Intel’s unit also found greater appetite for the technology among US-based respondents, with 40 per cent citing a desire to launch a GenAI system compared with 22 per cent from companies located elsewhere.

More than half of respondents from companies which have deployed GenAI reported improved customer experience, while 53 per cent stated GenAI provided better efficiency.

A majority (52 per cent) believe GenAI enhanced product capabilities and 47 per cent noted it had led to cost savings.

cnvrg.io stated slow adoption of LLMs by businesses was due to a lack of knowledge, along with cost and compliance concerns.

Some 84 per cent stated their skills need to improve due the increased interest in LLMs and 19 per cent that they had a strong understanding of how the models generate responses.

While the majority of organisations in the survey expressed a preference for building and customising their own LLM, 46 per cent viewed infrastructure as the biggest barrier for turning them into products.

Respondents also cited complexity and a lack of talent as additional barriers for AI adoption and acceptance.