Shadow AI is a leading indicator of what’s going to happen next
Steve puts his coffee down, wipes the cappucino froth from his beard, and rolls his eyes at the latest ask from upstairs.
“No worries. I’ll use ClaudeAI and smash this out in a few minutes. Maybe then go for a walk in the park with Azlan. It’s a lovely day.”
But his boss doesn’t know he’s using AI. Nor do Steve’s peers. Nor does the customer who’s data Steve is copying and pasting onto the internet. He’s logged on as steve@cappuccinobeard.com, not steve@acme.corp. He’s on the free plan.
What has Steve done wrong? Shall we pull the thread together?
“Shadow AI” is the unsanctioned use of AI by employees. It’s creating a new source of tension between employees and their employers. This article will examine the conflicting perspectives and key tradeoffs surrounding this emerging phenomenon.
The Employee’s Perspective: Creating shade
Shadow AI represents an opportunity to boost their individual productivity and performance. Employees turn to unapproved AI tools like chatbots or generative models to automate tasks, generate content, or accelerate their workflows – all without explicit permission from their organization.
Most companies don’t know how many SaaS apps they’ve got. They think “10ish” but the answer is, on average, 127. Now ask “How much AI is used here” — it’s the same vibe.
The employee rationale is simple – if using a certain AI system can save time and effort, why not leverage it, even if it exists in a gray area?
Whoa. Let’s put the brakes on. Employee enthusiasm for Shadow AI is often accompanied by concerns about job security. Steve might worry that by demonstrating the capabilities of AI, he could be gradually automating himself out of a role. What a quandary. There is a fear of “training the machine” to take over his responsibilities, potentially jeopardizing his long-term employment. And he’s just booked that Thailand trip. “Let’s keep this quiet”.
This dilemma puts workers in a difficult position: how does Steve weigh the short-term productivity gains against the potential long-term risks?
The Employer’s Perspective: Shining a light
From CxO Joan’s view in the rarified top-floor boardroom, Shadow AI represents a significant risk management challenge. She knows this because someone told her on the golf course yesterday. As she dreams through the steam off her cappuccino, looking down at the offices and workers below, she wonders:
How many of this lot are doing Shadow AI? What are they doing? How do I find out?
Joan is a leader. Her bonus and that ski trip to St Moritz next Jan depends on her ensuring the security and integrity of their data and systems, but the unsanctioned use of AI tools can introduce vulnerabilities. There are concerns about sensitive information being exposed through these unvetted channels, as well as the potential for non-compliance with industry regulations or internal policies. But how can you control what you can’t see?
One simple (!) and obvious thing that Joan and her company has to grapple with is the financial implications of Shadow AI. Unplanned and unbudgeted spending on AI tools — often paid for out of individual employee pockets — can represent a drain on company resources. This lack of centralized control and visibility makes it difficult for organizations to optimize their technology investments and align them with strategic priorities. In country vernacular: shit’s out of control.
There are other, worse, risks of Shadow AI, like information leakage. Picture the headlines.
Yet, at the same time, Joan is cognizant of the competitive landscape and the possibility that their rivals may be gaining an edge through the use of AI. There is a fear of falling behind the curve. Could this motivate a more permissive or even proactive stance towards AI adoption – even if it means ceding some control to individual employees? What if we just replaced employees? How would that impact my OKRs/KPIs and that St Moritz trip?
How to understand this, and how to deal with it?
Tensions and Tradeoffs
Is the fundamental tension at the heart of Shadow AI the balance between employee empowerment and organizational oversight? Is this two sides that don’t trust each other? Are we in the trench with our friend or our foe?
Employees seek the freedom to leverage AI tools that enhance their productivity, while employers strive to maintain control, security, and compliance. Right?
NOT ALWAYS.
Sometimes the employee is resistant to change and it’s the employer pushing for more use of AI.
This tug-of-war raises several key tradeoffs that organizations must navigate:
- Transparency vs. Secrecy: Should employers openly embrace and govern the use of AI, or should they crack down on unsanctioned activity to preserve control?
- Empowering Employees vs. Risking Job Displacement: How can organizations empower their workers to harness AI’s benefits without inadvertently automating them out of a job?
- Unlocking Efficiencies vs. Eroding Trust: Can the productivity gains of Shadow AI be balanced against the potential erosion of trust between employers and employees?
- Staying Competitive vs. Managing Change: Is the fear of falling behind the competition a sufficient justification for relaxing governance around AI usage?
Proposed Solutions and Best Practices
To address the challenges posed by Shadow AI, employers and employees must engage in open and constructive dialogue. Organizations should consider the following strategies.
Are these activities being budgeted for and are they measured and tracked?
- Establish clear policies and guidelines for the appropriate use of AI tools, striking a balance between employee autonomy and organizational oversight.
- Implement robust security measures to protect sensitive data and ensure compliance, while also making these tools readily available to employees through approved channels.
- Invest in reskilling and upskilling initiatives to help workers adapt to the evolving role of AI, preparing them for the changes ahead.
- Foster a culture of collaboration and shared responsibility, where employees are empowered to provide feedback and contribute to the development of AI-related strategies.
- Continuously monitor and analyze the use of AI within the organization, identifying opportunities to optimize workflows and mitigate emerging risks.
Steve roll his eyes again. He has no appetite for any of this. Shadow IT helps him get his job done, and he’s really enjoying the park. But what will happen next? Will he be here next year? Joan isn’t sure about Steve, or herself, or the St Moritz trip. And her coffee has gone cold.
