Join our daily and weekly newsletters to get the latest updates and exclusive content on leading AI developments.
A recent survey from cnvrg.io (an Intel company) shows that even though there’s a lot of excitement around artificial intelligence (AI), especially generative AI, not many businesses are actually using these technologies. The road to fully adopting these AI solutions is complicated due to various challenges like infrastructure and the lack of skilled personnel.
This survey, named the 2023 ML Insider and now in its third year, gathered insights from data scientists and AI professionals worldwide. It highlights a cautious approach towards embracing generative AI. Alarmingly, only 10% of organizations have managed to implement generative AI solutions in their operations, which is a small number considering the high expectations surrounding this technology.
Sectors like Financial Services, Banking, Defense, and Insurance are leading in adopting AI, benefiting from increased efficiency and better customer experience. However, Education, Automotive, and Telecommunications sectors are still very much in the early stages of AI adoption.
The survey points out that organizations might be slow to adopt generative AI because of challenges in implementing large language models (LLMs). Markus Flierl, corporate VP for the developer cloud at Intel, believes that with better access to affordable infrastructure and services like those from cnvrg.io and the Intel Developer Cloud, more organizations will adopt AI. This would make it easier to fine-tune, customize, and deploy existing LLMs without needing extensive AI expertise.
Additional insights from the survey reveal:
– 46% of respondents see infrastructure as the main barrier to deploying large language models for generative AI. These models demand heavy IT resources.
– 84% of those surveyed acknowledged the need to improve their skills to support the growing use of language models, with only 19% feeling fully proficient in using these models.
– The most common AI applications are chatbots and translation, reflecting generative AI advancements. However, only 25% of businesses have implemented any generative models in their operations.
– 58% of organizations have low AI integration with five or fewer AI models in use, a number that hasn’t grown much since 2022. Larger companies tend to use more than 50 models.
– 62% still find it challenging to successfully execute AI projects. The bigger the company, the harder it is to deploy AI.
Despite the buzz created by AI tools like ChatGPT, actual adoption in businesses is facing significant challenges. Companies are still testing generative AI rather than fully integrating it into their operations. Issues such as skill gaps, regulations, reliability, and infrastructure are major obstacles to rapid AI expansion.
Tony Mongkolsmai, a software architect and technical evangelist at Intel, emphasizes that the survey shows many AI developers believe a lack of technical skills is slowing down AI and ML adoption. He calls for the industry to reduce complexity and simplify tasks for developers to make AI more accessible.
To find out more, you can see the full ML Insider 2023 report on the company’s website.