How to Use AI to REDUCE EFFICIENCY and INCREASE Legal Costs

Legal AI

By Josh Stone, GC of Guardrail Technologies Inc.

The promise of artificial intelligence (AI) in business is to increase efficiency and reduce costs. In contract drafting and negotiation, large language models (LLMs) can produce initial drafts and analyze agreements in seconds—something that might take lawyers hours. Yet despite these advantages, AI often has the opposite effect if not used thoughtfully.

This article explores how misuse of AI in contract negotiations, can result in wasted time, increased legal bills, and even damaged professional relationships. Here’s how:

1. AI Can’t Read the Room

AI lacks human intuition. While seasoned negotiators can adapt based on expectations or interpersonal cues, AI cannot distinguish between a straightforward relationship and a more complicated negotiation. As a result, LLM-generated documents can misfire—providing content that is overly complex or too simplistic for the situation at hand.

Bottom line: When AI is used without human calibration and close review, it may produce output that is tone-deaf or mismatched to the context, sending the wrong signals to counterparties.

2. Misuse of AI Can Undermine Trust

Contracts and their negotiations are not just about legal terms—they are about relationships. Trust develops through forthright interactions, and parties form impressions based on communication style, fairness, and responsiveness.

Sending an AI-generated draft or commentary that’s not aligned with the stage, scope, tone, or substance of the negotiation can make the sender appear careless—or worse, disrespectful. AI that generates excessive or irrelevant comments may erode trust and derail negotiations.

Tip: Contract tone, balance, and complexity should be carefully managed by humans—not delegated to AI.

3. AI Wastes Time When Not Curated

AI can generate content quickly—but that doesn’t mean it's valuable. Long, verbose emails or sprawling document redlines can result from unedited AI output. Language generated by LLMs often bury key points in unnecessary fluff, creating confusion, wasted time, and frustration.

Real-world problem: I’ve frequently received AI-written emails that tried to sound impressive but buried the point in flowery language. Instead of helping, it made me think less of the sender.

In contracts, this effect is amplified. AI may flag issues as unfair just because they aren’t symmetrical—without recognizing differences in party roles or deal dynamics.

4. You Might Be Burning Billable Hours—Yours and Theirs

Bad AI output costs real money. Submitting a poor-quality draft to your own lawyer means they’ll spend more time fixing it—possibly at $500 to $3,000 per hour. The same goes for a counterparty’s counsel. When one party sends a document that isn’t fit for purpose, it can needlessly drive up everyone’s legal bills and sour the working relationship.

Advice: Don’t let AI create more work for people who charge by the hour.

5. Thoughtful Integration of AI Can Actually Reduce Costs

Despite the risks, AI can be helpful—if used wisely. Here’s how to integrate it productively:

  • Choose the right lawyer: Before asking your attorney to review AI-generated documents, ask whether they already have a solid starting form for the contract type. Using a forum they are comfortable with typically reduces drafting costs.

  • Use AI to assist, not micromanage: Don’t use AI to "second-guess" your lawyer or micromanage their process (e.g., asking a LLM to pick apart their work). Instead, collaborate with them on whether AI can streamline drafting and let them do their job.

  • Avoid unfair requests of professionals: Don’t ask an attorney to "bless" a document you generated with AI unless they’ve had meaningful input and are fairly compensated for risk and effort. As noted above, letting them run the process may actually reduce costs.

6. AI-Generated Comments Should Be Vetted, Not Forwarded

Never copy-paste AI-generated contract comments into an email without review. Sending a note like “please consider the points below” can lead your lawyer to overanalyze irrelevant comments, driving up costs.

Best practice: Vet every AI-generated suggestion:

  • Is it relevant?

  • Is it material?

  • Is it necessary in the context of the deal?

Only send the ones that truly merit discussion.

7. Avoid "Stump the Lawyer" Games and Similar Wastes of Time

Some employees misuse AI to challenge attorneys and impress management by using a LLM to spot a laundry list of potentially irrelevant issues. But playing "stump the lawyer" with obscure issues spotted by a LLM does more harm than good. Managers should ensure:

  • Employees understand how to use AI responsibly and not for ego or similar purposes

  • The company’s firms deliver real human expertise—not just AI drafting

  • Legal advice isn’t sourced blindly from machine-generated suggestions

Remember: You shouldn’t pay expert rates for AI-generated work product.

8. Know When to Intervene Like a Human

If you receive a draft or set of comments that’s clearly low quality, it’s okay to pick up the phone. A short conversation between humans can save days of unnecessary legal back-and-forth and the related costs. Depending upon the context, it may be reasonable to say something like:

“We looked at your draft, and it’s not really appropriate for this deal. Let’s recalibrate and save everyone some time and legal expense.”

Conclusion:
AI can reduce legal costs—but only when paired with human oversight, strategy, and judgment. If you want efficiency, not frustration, take the time to manage AI properly.

Josh Stone

General Counsel - Guardrail Technologies™

Josh has practiced corporate law for more than 25 years, focusing primarily on emerging companies, private investment funds and investment advisers. He has worked in various settings, including at leading international law firms, as an in-house general counsel at for-profit companies and nonprofit trade organizations and at his own entrepreneurial solo practice. Josh regularly handles private placements, venture capital investments, and corporate governance and securities work, as well as the formation of private investment funds. Josh has published in the area of AI, and is currently co-authoring an extended scholarly article on AI regulation along with a nationally recognized law professor.

Previous
Previous

10 Ways AI Is Accelerating Clinical Research (Without Cutting Corners)

Next
Next

Executive Order on AI Education: What It Means for Schools and the Law