You can decide your firm isn’t ready for AI yet. What you can’t decide is whether your paralegal already opened a chatbot in another tab this week to shorten a deposition summary or draft a client email. That decision got made without you, and the research says it almost certainly already has.
Microsoft’s Work Trend Index found that the large majority of knowledge workers now use AI at work, and more than half of them are using tools their employer never provided or approved. A 2026 workplace survey from Wakefield Research found two-thirds of office professionals had used AI even though they believed it wasn’t permitted under their company’s policy, and that number climbs higher at larger organizations. Nearly forty percent said, flatly, that they would rather keep using it quietly than risk being told to stop.
Legal has its own version of this gap. Industry surveys on gen AI adoption inside law firms keep turning up the same pattern: the staff most eager to use these tools, and often the first to actually use them, are the younger, more digitally native associates and paralegals, not the partners setting policy. One firm-level case study found a fifty-point gap between how many attorneys were interested in gen AI and how many were already using it, and the ones already using it were almost entirely junior, tech-comfortable staff acting on their own initiative.
That’s the part worth sitting with. The people closest to your case files, your client communications, and your discovery review are often the same people least likely to have been given any firm-sanctioned tool, any training on what’s safe to paste into it, or any guidance on where the ethical line sits. Left with a gap, they don’t wait. They open whatever’s free and familiar from their personal life and get the summary done.
Why younger staff reach for it first, and why that’s not a discipline problem
This isn’t about anyone being reckless. A recent enterprise survey found most office professionals believe they understand how to use AI for their job better than whoever is officially in charge of managing AI at their company. For staff who grew up with these tools already built into their phones and their coursework, that confidence is earned. It’s also exactly why silence from firm leadership reads as a vacuum rather than a boundary.
Separate research backs this up from the other direction. A global survey of enterprise workers found a meaningful share of employees using AI completely outside their IT department’s visibility, and many of those same employees reported their employer gave them no tools and no training at all, or training so irregular it didn’t count. The gap isn’t between firms that allow AI and firms that don’t. It’s between firms that have a stated position and firms that just haven’t said anything, which staff correctly read as permission by default.
What’s actually at stake in a law firm
For a paralegal, the ethical exposure is sharper than it is in most industries. AI can draft, summarize, and organize, but it cannot give legal advice, and it can produce confident, well-formatted, entirely wrong information. A paralegal using a consumer chatbot with no firm guidance is making judgment calls that should never fall on them alone: what case detail is safe to paste in, what counts as an unauthorized-practice line, what needs an attorney’s eyes before it goes anywhere near a client. Without training, without a sanctioned tool, and without a clear policy, that judgment call happens anyway. It just happens invisibly, on a personal account, with no record of what went in or what came out.
That’s the actual risk. Not that staff are using AI. That they’re using it without anyone at the firm having decided what “using it well” looks like.
The fix isn’t a ban
Firms that try to shut this down outright tend to get exactly the outcome the research predicts: staff use it anyway, just more quietly. One 2026 study on unapproved workplace AI use found that once employees are given a company-sanctioned system that actually works for them, the incentive to go looking for an unofficial alternative disappears. The fix isn’t a memo telling people to stop. It’s giving them something better than what they’re already using, with the training and the ethical guardrails built in from the start.
That’s a deliberately smaller ask than most firms assume. It doesn’t require a firm-wide AI strategy or a six-month rollout. It requires picking one sanctioned tool, teaching staff what it’s for and where its limits are, and saying out loud that this is the approved path instead of leaving everyone to guess.
If your firm hasn’t made that decision yet, your staff already has, on your behalf, without you in the room.
What to do this week
Ask your paralegals and support staff, directly and without judgment, what they’re already using. You’ll likely learn more from that one conversation than from any policy document. Then decide, together, what the firm’s actual position is, and put it in writing where everyone can see it.
AI built for law firms. Powered by real lawyers.
Book a Free Exploration Call · See a Working Demo · Browse the AI Systems Catalog
Leave a Reply