Shadow AI is on the rise and numerous case studies have shown that it can be a major issue for organizations that don't have comprehensive AI governance in place. Discover how SAP LeanIX can empower you to leverage AI without the risk.
"Shadow artificial intelligence (AI)" is a natural consequence of excitement about a revolutionary technology in a business-led IT environment. AI sounds like magic, so of course your employees want to try it out in their work.
Not only does this leave you open to a variety of security and operational risks, but it also means you might lose touch with what your employees are doing with AI. To keep oversight of your AI landscape and avoid AI risk, you need to have a proper AI governance framework in place.
Our AI survey, however, shows that only 19% of organizations have such a framework in place. It's clear that companies need to get oversight of AI, and they need to do it soon.
To find out more about the results of our AI survey, download our report:
DOWNLOAD: SAP LeanIX AI Survey Results 2024
What Is Business-led IT?
'Business-led IT' means allowing the needs and preferences of the employees within your organization to guide you in IT provisioning. There are tremendous advantages to allowing your employees to make decisions regarding their IT toolset.
Your sales team is more likely to know whether Salesforce, Seismic, or Zendesk is best for your organization than an IT team that has never used them. Likewise, with the rise of software-as-a-service (SaaS) applications, non-technical users, such as sales specialists, can often implement and maintain these applications without IT support.
IT teams and enterprise architects, however, will have a better understanding of your organization's budget and the compatibility of different platforms with the rest of your IT landscape. Not to mention, it's easy for different teams to spend money on two instances of the same platform, when a single instance could improve collaboration and save costs.
Business-led IT, therefore, doesn't mean just allowing your employees to run your IT landscape however they wish. It means, rather, that enterprise architects and IT teams should control your IT infrastructure, but ensure that fulfilling the needs of users and empowering them with self-service capabilities is your priority.
This is the difference between a successful business-led IT methodology and the well-documented issues that come with 'shadow IT'. This sinister-sounding concept is what happens when employees are permitted to leverage whatever tools they think best without IT oversight.
The result of shadow IT is an inefficient, unnecessarily expensive, chaotic IT landscape that serves the needs of some individual users in their particular day-to-day work, but doesn't serve your business overall. With the rise of artificial intelligence (AI), 'shadow AI' is a dangerous sub-category of shadow IT that is causing issues for IT teams.
What Is Shadow AI?
'Shadow artificial intelligence (AI)' is the unsupervised use of AI tools within your organization that can cause operational, legislative, and security issues. Without oversight and an AI governance framework, well-meaning employees attempting to leverage AI tools improperly can cause disaster.
AI regulation is rapidly being enacted across the world, AI is a new technology that is still little understood, and employees have no experience or reference in the use of the new technology. It's, therefore, dangerous to allow AI to be leveraged without supervision.
Your colleagues are, naturally, excited to try out this revolutionary technology and to leverage its capabilities to enhance their productivity. Without proper guidelines in place, shadow AI will run rife.
Shadow AI is a natural consequence of business-led IT, but that doesn't mean taking the IT needs of your business into account is the wrong approach. IT teams should make meeting the demand for AI tools a priority, but only within a governed framework.
AI governance, therefore, actually empowers your organization to fully leverage the power of AI tools without risk. Meanwhile, unsupervised AI use will eventually need to be prohibited unless it can be controlled.
The Risks Of Shadow AI
Shadow artificial intelligence (AI) can cause tremendous real-world consequences without the supervision of a proper AI governance framework. There have been numerous recent cases of unrestricted AI experimentation causing issues for organizations:
Samsung Publishes Proprietary Code
Samsung was forced to ban its employees from using ChatGPT recently. This was because it found employees were uploading proprietary code into the large language model (LLM) in order to find and fix bugs.
Air Canada's "Lying" AI Chatbot
Air Canada's AI chatbot recently told a customer that they could apply for a partial refund of the cost of their travel due to a bereavement. This was, however, in contradiction to Air Canada's bereavement policy, though a court upheld the chatbot's error, treating it as an employee of the airline.
ChatGPT Fabricates Legal Precedent
Attorney Steven Schwartz attempted to use ChatGPT to do legal research to support his case against Colombian airline Avianca. Unfortunately, all the precedents ChatGPT found were invented by the LLM, and the case was dismissed and Schwartz fired.
Zillow Bets On AI And Loses $300m
Real estate company Zillow hoped a machine learning algorithm could accurately guess the price of properties for a quick sale. However, the algorithm turned out to be wrong in 2-7% of cases, causing Zillow to write off US $300 million and 2,000 employees redundant.
Amazon's Recruitment AI Won't Hire Women
Amazon secretly developed an algorithm to screen resumes to help it hire key talent. However, since the AI was trained on historical recruitment policies, it was found to be screening out womens' CVs and was subsequently retired.
Bringing Shadow AI Into The Light
Shadow artificial intelligence (AI) is nowhere near as sinister as it sounds, so long as you have proper oversight and governance of business-led AI use in your organization. Implementing an AI governance framework, however, requires documentation of your AI landscape.
SAP LeanIX is a repository for fact sheets regarding all the applications and IT components in use in your organization. This includes information about artificial intelligence capabilities and how they support your essential business capabilities.
This information can then be visualized and shared with your stakeholders to make sure everyone is on the same page when it comes to AI. AI documentation enables AI governance, which enables AI adoption.
To find out more about how SAP LeanIX can support AI governance and adoption, book a demo:
