Your Employees’ Hidden AI Habits That Put Your Business at Risk

Today's employees are stretched thin. They're juggling 15 Slack channels, 300 unread emails, back-to-back meetings, and hair-pulling deadlines. When they discover that AI can summarize reports, draft emails, and generate campaign ideas in seconds, their feelings of relief are palpable.

The problem is, they’re often using these AI tools without company approval or oversight. According to new Upwork research, 55% of full-time employees admit to using unapproved AI tools at work.

This quiet, often under-reported use is known as shadow AI, and it can create serious risks for your business.

Hear the full conversation in Work Week

Read transcript

How everyday work pressures lead to unsafe habits

Employees who turn to unapproved use of AI tools aren’t being rebellious. They’re drowning in their workloads and they see AI tools as a life raft. What many don’t realize is their life raft might have a hole that’s quietly leaking your business’s most sensitive data.

How is this possible? When a user gives an AI tool a prompt, the AI often stores any data the prompt contains. Later, this data can be used to train future AI models. When users then ask these new models the right questions, your private info can unintentionally become public.

Here's how easily that could happen:

• Your marketing manager pastes your unreleased product roadmap into ChatGPT with a prompt to make the language more exciting.
• Your finance lead uses an AI code assistant to automate error checks in their weekly report.
• Your service agent feeds customer emails into a writing assistant to draft faster responses.

Even a single "harmless" prompt can spark serious consequences including compliance violations, leaked client data, and reputational damage. 

Shadow AI is becoming such a problem that, according to IBM’s Cost of a Data Breach Report 2025, it has been the cause of a security breach for 1 in 5 companies, and the average cost per breach is $4.4 million. Moreover, Gartner estimates that shadow AI will cause security and compliance incidents for over 40% of global organizations by 2030. The cost per incident, and the scope of the issue, should make any company take note, and be a particular concern for small and medium-sized businesses (SMBs).

Why bans can backfire

For leadership, a natural reaction may be to ban AI tools. However, that 55% of employees admit to already using shadow AI tools suggests that an outright ban might simply be ignored. And  according to IBM, 28% of employees explicitly stated they’d keep using AI even if it were forbidden. 

In sum, banning AI use may create more secrecy rather than more security.

Keep in mind that your employees’ intentions are good; it’s their execution that’s risky. And that’s not entirely their fault.

The same IBM report shows most businesses lack proper AI oversight and guidance:

• 97% of organizations that experienced AI-related security incidents lacked proper access controls
• 63% had no governance policies in place to manage AI use or detect unauthorized tools

Moreover, the research found that 68% of full-time employees want more support from their organization to experiment with AI tools.

These are people looking for solutions and finding on the one hand a lack of organizational guidance, and on the other an abundance of AI tools on the market.

Don’t worry, you can fix this

The employees testing AI tools on their own are innovative and curious. And our research shows that this is a majority of employees; 78% of full-time employees in our survey educate themselves about AI, outside of their organization, when they aren’t getting enough through work.

Innovation, curiosity, and the drive for self-improvement are all valuable traits. Instead of shutting them down, consider channeling them safely.

Start with the right approach

Before you write a single policy, shift how you think about the use of artificial intelligence at work. Don’t treat AI use as though it’s rule breaking. Your employees aren't being sneaky; they're being creative. So work with them to set boundaries.

Tap into their curiosity

If someone discovers a helpful AI tool, create a culture that encourages them to tell you about it, not hide it. Set up approved ways for people to test new tools and share what they learn. That's how innovation happens.

Find out what's really happening

Send a quick, anonymous survey to your team. Ask:

• What AI tools are you using right now?
• What do you use them for?
• What tools do you wish you had?

Tell them up front that you’re not policing them; you’re trying to better support them. You might be surprised by what you learn. Some employees might already be AI power users. Others might have creative productivity hacks you can safely scale across the business.

Create a simple AI guide

Balance experimentation with safety through guardrails. Create a one-page "AI Quick Guide" that covers:

• Tools you've approved and why they're safe
• Information they should never put in AI (e.g., client data, financials, passwords, unreleased products)
• How to request and test a new tool
• Who to contact with questions

Post the guide where everyone can easily find it. Update it as you learn.

Bring in people who've solved this before

If you don't have the time or in-house expertise to tackle this, consider bringing in freelance AI specialists to help. The new research shows that freelancers are more likely than their full-time employee counterparts to cite having “extensive experience” in some key areas of AI use and development. These include building and training AI models, data science, and chatbot development.

By accessing this skilled talent on-demand, you can better support your employees in their search for AI tools, without compromising your organization’s security. 

How AI-smart freelancers deliver better work, faster

Huntr once hired freelance marketer Ashliana Spence for a keyword project they estimated would take a week. She used AI to automate part of the process and finished the work in a fraction of the time, which allowed her to start their other projects sooner. By combining her marketing expertise with the right AI tools, she delivered high-quality work across every task.

“She brings fresh ideas, improves ROI, and helps us get more value out of every campaign,” said Sam Wright, Head of Operations and Partnerships at Huntr.

See the full case study.

Keep the conversation going

Set up a Slack or Microsoft Teams channel where people can share what's working, ask questions, and trade tips. You could also have a power user host short monthly gatherings where anyone can join and explore new tools and use cases together.

Make AI your advantage, not your risk

Your employees are going to use AI tools. The real choice is whether they use them in the shadows or out in the open.

When you encourage responsible use and put simple guardrails in place, AI shifts from a potential risk to a shared engine for business growth. That’s when teams start to see real benefits.

And if you’d like support, Upwork is ready to help. Check out the world’s largest marketplace of freelance AI specialists.

Research note: This article includes findings from the Upwork Research Institute’s 2026 Future Workforce Index survey of 2,400 U.S.-based workers (March-April 2026).

Upwork is not affiliated with and does not sponsor or endorse any of the tools or services discussed in this article. These tools and services are provided only as potential options, and each reader and company should take the time needed to adequately analyze and determine the tools or services that would best fit their specific needs and situation.