Shadow AI – hidden impact of deploying Artificial Intelligence
The other day I met with a client of mine, and the conversation naturally fell on the impact of artificial intelligence. The central part of the discussion focused on the state of AI model implementation in organizations today.
Several studies suggest that the average number of models deployed today is around five, but that growth rate over the next four years will be slightly above 60% CAGR. In four years, the average large organization will have 35 models deployed, broadening the impact of artificial intelligence. However, this is far from the full effect. Consideration will also need to include several other elements impacting organizations.
The number of models deployed in surveys by Gartner and McKinsey, only consider officially known artificial intelligence models. It is highly plausible that every organization will see a rise in the deployment of “Shadow AI” models. No different than the historical proliferation of Shadow, originally meaning end-user purchased PCs and software. The increase in artificial intelligence will cause a similar trend. As functional departments like Marketing, Supply Chain, and HR increasingly decide to deploy solutions, including AI models, themselves. The broadening deployment of AI will result in the proliferation of a shadow AI model environment.
But this is not the only impact. As models are deployed and continuously improved, their complexity will increase. Complexity means the complexity of the task or the need to be handled by the model. A model will likely also increase in breadth, handling more tasks done by a single person or a product or service. And lastly, it is highly likely that an AI model deployed in 2020 will see an increased number of users annually. As models mature, it may become more economically and technically viable to deploy models to more users within a function (like marketing or HR) or even increase the number of business functions that implement the model.
Managing AI across an organization is not solely about an increased number of models deployed over the next few years. Managing AI is equally about managing the complexity of models, breadth models, the number of users, and the proliferation of Shadow AI. This proliferation of AI and Shadow AI requires an enterprise approach to managing AI and a well-defined AI Governance strategy. With time artificial intelligence will require more complex procedures and governance.
So, what are three things that the organizations need to consider?
How will they manage an increased number of AI models deployed across a wider part of the organization? And as a result of that, how will the address a likely rise in Shadow AI?
What is AI governance to the organization, who is accountable for AI governance, and responsibility do they have?
How does the organization communicate about AI, the accountabilities everyone in the organizations, the requirement for all to learn skills around AI and lastly to understand how to deal with Shadow AI.
The complexity of artificial intelligence goes far beyond an increase in the number of models deployed. CxOs of organizations needs to be particularly focused on the potential impact of Shadow AI.
Source: The Sondergaard Group
About the author
Chairman of the Board, 2021.AI
Peter Sondergaard is currently Chairman of the Board at 2021.AI and Owner of his Executive Advisory company, the Sondergaard Group. Before this, Peter worked as the Executive VP and member of Gartner’s operating committee for 15 years. Peter is a well known and sought out speaker covering many topics within IT, AI & ML.
You might also like…
Putting AI models in production is notoriously tricky, and the challenges have many different nuances. This blog post summarizes the fundamental difficulties of productionizing AI models and how we, at...
The Ethical AI Newsletter
It’s not fake. It’s not artificial. It’s real news! Sign up for our Ethical AI newsletter and get the latest AI insights from our data science and AI experts.