Results from the Roadblocks to Scale survey indicate that while there is still work to be done, advances in data discovery and management, skills training and AI explainability are driving the rate of AI adoption faster than many predicted.
For example, 45% of the respondents from large companies (1,000+ employees) said they have adopted AI, while 29% of small and medium sized businesses (under 1,000 employees) responding said they have adopted the technology. These numbers are significantly higher than some industry watchers have estimated to date. Some of the more telling data points of the survey, for which the Executive Summary can be found here, include the following:
- Skills. Major roadblocks are still holding companies back from the benefits of AI. Amongst respondents, 37% cite limited AI expertise or knowledge as a hinderance from successful AI adoption at their business, with increasing data complexities and siloed data (31%) and lack of tools for developing AI models (26%) following close behind.
- Trust. Trust is part of the bedrock of AI’s deployment. Globally, 78% of respondents across all countries surveyed say it is very or critically important that they can trust that their AI’s output is fair, safe, and reliable. Moreover, being able to explain how AI arrived at a decision is universally important (83% of global respondents).
- Data. Companies currently deploying AI technologies are more likely to use a hybrid cloud (38% adopted) or hybrid multicloud (17% adopted), as AI success is fed by data. And, data is everywhere, on all clouds.
Based on our interactions and the results of this study, we expect to see organizations not only adopt AI – but scale it across their enterprises, by building/developing their own AI, or putting ready-made AI applications to work. For example, according to the survey, 40% of respondents currently deploying AI said they are developing proof-of-concepts for specific AI-based or AI-assisted projects, and 40% are using pre-built AI applications, such as chatbots and virtual agents.
I see the excitement building with clients every day. Consider just a couple of recent examples. Legal software developer LegalMation has leveraged IBM Watson and our natural language processing technology to help attorneys automate some of the most mundane litigation tasks, speeding, for example, the written discovery process from multiple hours to a few minutes.
Wunderman Thompson Data, the global digital agency, uses Watson Studio and Watson Machine Learning to access, analyze and run models on terabytes of data stored across its hybrid multicloud environment. As part of its work with Watson Machine Learning, Wunderman is also using IBM AutoAI to automate models to analyze literally tens of thousands of features from across the company’s datasets.
When I look at insights from the report, which was conducted by the firm Morning Consult, and think about my interactions with clients, the roadblocks to AI adoption have been a prime concern. They’re the reason we’ve worked to lower the barriers of entry and make AI more accessible to businesses.
It’s why we launched the Data Science Elite Team in 2018, a global group of experienced technical professionals who help companies solve real problems with AI. It’s what drove us to introduce innovations like Watson OpenScale, to help mitigate bias in AI models; Watson AutoAI, which literally uses AI to build AI models; and it’s what led us to create the first-of-its-kind container-based data analytics platform, Cloud Pak for Data, that lets people run Watson on any cloud. We’ve also, taken skills training and support to whole new levels, with robust data science work with several open standards bodies, such as The Open Group and The Linux Foundation.
2019 was a productive year for AI, but 2020 is shaping up to bring an exciting new level of commitment and with it, exciting new outcomes for all involved.