Study Reveals AI ‘Revolution’ Progressing Slowly in Enterprises

A recent survey by cnvrg.io, part of Intel, highlights the slow enterprise adoption of artificial intelligence (AI) solutions, despite the widespread excitement surrounding generative AI. This 2023 ML Insider survey, now in its third year, reveals ongoing challenges in moving generative AI solutions to production, including infrastructure issues and skill shortages.

Key findings indicate that only 10% of organizations have successfully deployed generative AI solutions. Financial services, banking, defense, and insurance sectors are leading the way, harnessing AI to enhance efficiency and customer experiences. In contrast, industries like education, automotive, and telecommunications remain cautious, with many AI initiatives still in the early stages.

Markus Flierl, Corporate VP for the Developer Cloud at Intel, notes, “Organizations may hesitate to adopt generative AI due to barriers in implementing large language models (LLMs). However, increased access to cost-effective infrastructure, such as that offered by cnvrg.io and Intel Developer Cloud, could facilitate greater adoption by making the customization and deployment of LLMs more manageable without requiring extensive AI talent.”

Additional insights from the survey include:

- 46% of respondents identified infrastructure as the primary barrier to deploying large language models, which can be resource-intensive.

- 84% acknowledged a need for improved skills to support the rising interest in language models, with only 19% feeling fully proficient in content generation via these models.

- Top use cases for AI included chatbots and translation, reflecting generative AI advancements in 2023, yet only 25% of organizations have deployed any generative models.

- 58% reported low AI integration, utilizing five or fewer models, a statistic that has seen little change since 2022; larger companies are more likely to deploy over 50 models.

- 62% of organizations find executing successful AI projects challenging, with larger companies facing greater difficulties in AI deployment.

These results underscore that, despite the buzz from tools like ChatGPT, enterprise-level AI adoption is hindered by tangible obstacles. Companies tend to experiment with generative AI rather than fully integrating it into their operations, facing challenges such as skills deficits, regulatory constraints, reliability issues, and infrastructure limitations.

Tony Mongkolsmai, software architect and technical evangelist at Intel, states, “The 2023 ML Insider Survey shows that a majority of AI developers believe a lack of technical skills is impeding their organization’s adoption of machine learning and large language models. As an industry, we must focus on simplifying processes and reducing complexity to support developers in this fast-evolving landscape.”

For more insights, access the full ML Insider 2023 report on the company website.