AI Settlement Ethics: Lawyers’ 2026 Disclosure Dilemma

A new study from the American Bar Association (ABA) found that 37% of lawyers have serious ethical concerns about AI settlement tools, especially around what we’re supposed to tell our clients. That number gets right to the heart of the problem: these AI platforms promise incredible efficiency, but our ethical duty to keep clients fully informed hasn’t changed a bit. We have to figure out how to integrate AI into our negotiation strategies without breaking the trust that our clients place in us.

Data Point 1: 42% of personal injury firms in Georgia now use AI for case valuation or settlement prediction

That 42% figure, pulled from a 2025 Georgia Bar Association survey, shows just how fast AI is taking over, even in a niche like personal injury. On the ground, this means if your firm isn’t at least looking into tools like Everlaw for discovery or Casetext’s CoCounsel for research, you’re already falling behind your peers in the state. The pressure to use this tech for better efficiency and more accurate settlement predictions is intense, but this rapid adoption is also forcing a conversation about ethics. When a lawyer uses an AI model to predict the settlement range for a client’s car accident claim, for example, the client has a right to know a machine helped generate that number. Failing to disclose that is a breach of trust waiting to happen, especially if the client assumes the advice came purely from your human expertise.

Data Point 2: Only 18% of clients fully understand how AI impacts their legal case, even when informed

A recent report on consumer legal services from the National Center for State Courts shows a massive gap between lawyers disclosing AI use and clients actually understanding it. A comprehension rate of just 18% is a major red flag. Simply saying, “we use AI” is totally insufficient. We have a duty under Georgia Rule 1.4 (Communication) to explain what this means in plain language for their specific case. Think about a workers’ compensation claim where an AI analyzed hundreds of similar Fulton County Superior Court cases to suggest a settlement amount. The client has to understand that the AI is a predictive tool working off historical data. It’s not a crystal ball. They need to know its limits, like its inability to factor in the unique emotional pain of their injury or the fact that a specific opposing counsel is known to be a weak negotiator. Without that clarity, clients might give an AI’s projection far too much weight, which can easily lead to a breakdown in the relationship or even a claim of inadequate representation later on.

Data Point 3: Algorithmic bias in settlement predictions led to a 15% disparity in offers for minority clients in a simulated study

This simulation from researchers at the University of Georgia School of Law should give every practicing attorney pause, as it points to a terrifying ethical pitfall. While it was a simulation, it reveals how these systems can go wrong. The AI models are trained on historical data, and that data is often polluted with the societal biases of past legal outcomes. If past settlements for certain demographics were systemically lower, an AI trained on that data will just perpetuate the same unfairness. This is a huge issue for Georgia personal injury lawyers with clients in diverse areas, from neighborhoods in Buckhead to rural communities around Gainesville. Can you imagine an AI suggesting a lower settlement for your client injured in a truck accident on I-75 just because its training data showed lower average payouts for people with a similar demographic profile, completely independent of the facts of their case? This is an ethical disaster and a fast track to being accused of discrimination. We have to diligently check the AI tools we use, asking hard questions about their data sources and understanding their potential for bias before ever applying them to a client’s case. Our professional judgment must always override the algorithm.

Data Point 4: Firms with clear AI disclosure policies report a 25% higher client satisfaction rate in AI-assisted cases

Here’s the business argument for doing the right thing. A legal tech survey from the American Bar Association found that proactive ethical disclosure pays off, showing that firms with clear AI policies have a 25% higher client satisfaction rate. Clear policies build trust. When clients get a sense of how AI is being used (even without a computer science degree), they feel more respected and involved. In a complex medical malpractice case in DeKalb County Superior Court, explaining how AI is helping you analyze thousands of pages of medical records or predict jury tendencies demystifies the process. It makes AI a valuable tool working for them. A good policy might just be an addendum to the retainer agreement that names the tools you use, like Relativity Trace for analyzing communications or Westlaw Precision for its predictive analytics. Explaining that these things help find patterns and assess risk, but that you, the lawyer, make the final calls, manages expectations and reinforces the attorney-client relationship.

Challenging the Conventional Wisdom: “AI is just another tool, like legal research software.”

I keep hearing lawyers shrug off the ethical worries around AI, saying it’s no different than Westlaw or LexisNexis. I fundamentally disagree. Those platforms made research more efficient, but they didn’t generate settlement recommendations or assess case values with their own algorithms. Traditional research tools are passive databases. You search, they find. The AI we’re dealing with now, especially generative and predictive AI, is active. It processes, interprets, and synthesizes information in a way that mimics human reasoning, just without any of the attached ethical grounding or conscience. This difference is everything when it comes to client disclosure. No one ever expected you to tell a client, “I used books.” You don’t disclose, “I used a database.” But when an AI system ingests a client’s entire personal injury file from a wreck on I-285, medical records, police reports, witness statements, to then *suggest a settlement figure*, that’s a different world. The AI is forming an opinion, a probabilistic assessment that directly influences your advice. This requires more disclosure and a real discussion about the tool’s limits. Ignoring this difference puts our fiduciary duty to our clients at risk.

AI is going to be part of settlement negotiations, period. But deploying it ethically requires constant vigilance and transparency. We must proactively educate our clients on AI’s role, making sure they understand its benefits and its real-world limitations. By putting clear disclosure first and always maintaining human oversight, we can use AI’s power without abandoning our core professional responsibilities. For more on what to expect, see our article on Personal Injury AI: 30-40% Efficiency by 2027.