As the increasing trend of artificial intelligence (AI) incorporation continues, new research underscores a seemingly paradoxical trend. The 2023 ML Insider survey conducted by cnvrg.io, an Intel company specialising in AI and large language model (LLM) platforms, revealed that only 10% of organisations have integrated generative AI (GenAI) solutions into production this past year.
The annual survey offers a comprehensive analysis of the machine learning industry and gleans invaluable insights from technology professionals worldwide. This year's study, the third of its kind, surveyed a global panel of 430 professionals to understand better their developmental strides in AI and how generative AI is being applied in their businesses.
Despite generative AI rattling the technology industry conversation in 2023, the survey reveals a hesitancy amongst the majority of organisations to adopt the technology. This reluctance could arise from challenges organisations face during the implementation of LLMs. Markus Flierl, Corporate Vice President and General Manager of Intel Cloud Services, shared his insights on the situation.
"Despite the interest, organisations seem hesitant to adopt generative AI potentially due to barriers of implementing large language models," said Flierl. He added that "with greater access to cost-effective infrastructure and services, like those offered by cnvrg.io and the Intel Developer Cloud, we expect greater adoption in the next year as it will be easier to fine-tune, customise, and deploy existing large language models without requiring AI talent."
In addition to the only 10% of organisations indicating they have launched GenAI solutions to production in the past year, the survey highlights a seemingly low adoption rate of GenAI technology within businesses. However, amongst those organisations implementing GenAI models, about half report improved customer experiences (58%), enhanced efficiency (53%), upgraded product capabilities (52%), and significant cost savings (47%).
Despite the tangible benefits, the path to GenAI adoption is not without challenges. The survey reveals that 46% of respondents view infrastructure as the greatest hindrance to developing LLMs into products. Other obstacles flagged by respondents include complexity, lack of AI talent, high cost of implementation and compliance issues. Of the respondents, 84% stated they need to enhance their skills due to mounting interest in LLM technology, with only 19% confident in their current understanding of this technology's response mechanisms.
On a positive note, the survey indicates that GenAI technology is having a significant industry impact. Compared to 2022, the application of chatbots/virtual agents has jumped by 26%, and translation/text generation has increased by 12% in 2023, reflecting the advances in LLM technology. Those organisations that have already deployed GenAI speak of multiple benefits derived from the application of LLMs, including enhanced customer experience and efficiency, improved product capabilities, and cost savings.