AI tools such as ChatGPT can provide quick and confident answers to legal and business questions. But AI-generated responses are not a substitute for legal advice. AI may recommend a strategy without understanding the governing agreements, the full factual record, or the legal consequences of the proposed conduct. Using AI in a legally sensitive matter can also create risks involving confidentiality, evidence, and discovery.
The Delaware Court of Chancery’s decision in Fortis Advisors LLC v. Krafton, Inc., 354 A.3d 906 (Del. Ch. 2026), illustrates two important risks: AI can recommend a legally damaging strategy, and AI conversations may later become evidence.
AI Can Recommend a Legally Damaging Strategy
In Fortis Advisors, a corporate executive used ChatGPT while considering how to address substantial earnout obligations arising from an acquisition. After ChatGPT indicated that canceling the earnout would be difficult, the executive asked what the buyer could do if it could not reach an agreement with the acquired company’s executives.
The resulting AI-generated plan proposed a pressure campaign involving operational controls, legal-defense preparations, public messaging, and aggressive and conciliatory tactics intended to force a compromise.
The problem was that the strategy focused on creating leverage rather than determining whether the proposed conduct was permitted by the acquisition agreement. It did not adequately account for contractual protections or the potential consequences of interfering with those rights.
The court ultimately found that the buyer breached the acquisition agreement and that its stated reasons for terminating the executives were pretextual. The AI-generated strategy became evidence relevant to the buyer’s intentions and conduct.
The lesson is important: AI can suggest a way to achieve a desired result without adequately considering whether the result—or the means used to obtain it—is legally permissible. A confident AI-generated response may fail to account for the specific contracts, facts, rights, and potential liabilities involved.
AI Prompts May Be Discoverable
Businesses and individuals should not assume that conversations with ChatGPT or another AI tool are confidential or protected by attorney-client privilege. An AI conversation should not automatically be treated as a privileged legal communication.
AI prompts can reveal:
- The result a client is attempting to achieve;
- An obligation the client wants to avoid;
- Concerns about whether a justification is legitimate;
- Anticipated claims or defenses; and
- The motivation for taking a particular action.
In Fortis Advisors, the AI-generated plan became part of the evidentiary record, and the court noted that relevant ChatGPT conversations had been deleted.
Depending on the circumstances, AI prompts, responses, screenshots, exports, and internally circulated AI-generated documents may be subject to preservation and discovery. AI-generated content may also become relevant evidence concerning a party’s intentions, strategy, or conduct. Deleting potentially relevant AI records after a dispute arises can create additional legal risks.
Practical Guidance for Businesses Using AI
Before using AI for a legally sensitive matter, businesses should:
- Consult counsel before acting on an AI-generated legal strategy.
- Verify AI-generated information rather than assuming a confident response is legally correct.
- Assume AI conversations could become evidence if they are relevant to a dispute or litigation.
- Protect confidential information by avoiding privileged, confidential, or sensitive information in unapproved AI tools.
- Preserve relevant AI records when litigation is pending or reasonably anticipated.
AI can be a useful informational tool, but it is not a substitute for qualified legal advice. For matters involving legal rights, contractual obligations, disputes, or potential liability, businesses should consult an attorney before relying on or acting upon AI-generated guidance.
About Us:
Newman & Lickstein, LLP advises businesses and organizations on e-discovery obligations, litigation risk management, data preservation, and complex commercial litigation matters. Our attorneys work closely with clients to address evolving legal and technological issues involving electronically stored information, cloud-based systems, and defensible preservation practices in modern litigation.
