A recent O’Reilly survey reveals that more than 80% of enterprises are using AI in deployed applications.
The recent survey from O’Reilly says that over half of enterprises are currently in the mature phase of artificial intelligence (AI) adoption. This indicates there has been significant growth within the past year. O’Reilly’s AI Adoption in the Enterprise 2020 report revealed 85% of companies are evaluating or using AI in production. Over half of the organizations surveyed said they are mature adopters of AI or are using AI for analysis or in production.
As compared to O’Reilly’s 2019 report, only 27% of companies were in the phase of mature adoption, while 54% were in the evaluation phase. Nearly half of the respondents cited that the majority of AI tech is being used in research and development, followed by IT, customer service, and marketing, advertising, and PR. Giving insights about the challenges in AI adoption, the survey said the 22% of respondents still believe lack of institutional support is the biggest obstacle in 2020 the same as in 2019. Identifying appropriate business use cases (20%), skills gap (17%), and data quality issues (16%) were some of the other obstacles.
The shortage of AI and machine learning (ML) skills is another consistent struggle, found the survey. Nearly 60% of respondents said the shortage of ML modelers and data scientists topped the list as having the biggest skills gap. The same skills areas were issues in 2019 as well, mainly because the most critical skill shortages cannot easily be addressed.
According to 66% of mature AI practitioners, unexpected outcomes were the single most common risk factor at the time of building and deploying ML models. Mature adopters who were surveyed cited the need to control for interpretability and transparency of ML models as the second-most cited risk factor. On the other hand, respondents who are still evaluating AI noted fairness, bias, and ethics in second place.
The topmost popular ML technique for all adopters was supervised learning. Meanwhile, mature adopters mentioned deep learning was the second most used technology. Enterprises in the evaluation phase said deep learning is slightly more used than supervised learning.
AI adoption is multiplying at a slow and steady pace. The majority of enterprises that were experimenting with AI are now using it in production deployments. Despite the fact that almost half the enterprises are not there yet, companies need to go the extra mile to put their AI efforts on solid ground. It is encouraging that there are a sizeable number of companies that have reached the mature stage and that is going to be the driver for all the benefits that’s AI provides to enterprises.