Delaware’s 2026 DGCL Overhaul: What Section 144, Section 220 and AI Chat Discovery Mean for Boards

Delaware corporate law and the discoverability of AI conversations might seem like unrelated topics, but for 2026 boards they are converging into a single governance problem: how directors and officers document, and defend, the decisions they make with the help of new tools and under new statutory rules. Delaware’s amended Section 144 and Section 220 reshape the legal terrain for conflicted transactions and stockholder inspection, while a fast-moving set of court decisions is now defining whether an executive’s private conversation with an AI legal assistant can be pulled into evidence.

What Changed Under Delaware’s Amended Section 144?

Amended Section 144 creates statutory safe harbors for transactions involving director conflicts of interest or controlling stockholders, replacing a more litigation-prone common-law standard with clearer, codified approval pathways. Enacted through Senate Bill 21 in spring 2025 and now shaping 2026 practice, the amendment gives boards a defined process for insulating conflicted transactions from later challenge, building on the same Section 144 independence standard that redefined director conflicts earlier in 2026.

For directors, the practical effect is more predictability: following the statutory safe harbor procedures — such as approval by a disinterested and informed board majority or a fully informed stockholder vote — provides a much stronger defense against breach-of-duty claims than relying on general fairness arguments after the fact.

How Does Amended Section 220 Limit Stockholder Inspection Rights?

Amended Section 220 narrows the scope of books-and-records demands that stockholders can make of Delaware corporations, curbing what had become an increasingly broad discovery tool used to build litigation before a formal lawsuit was even filed. The amendment was a direct legislative response to years of expanding Section 220 demands that companies argued had turned into a pre-litigation fishing expedition.

Boards should treat the narrower standard as a reduction in exposure, not an elimination of it — well-documented, properly minuted board processes remain the best defense, since Section 220 still permits access to records reasonably related to a stockholder’s stated, proper purpose. That defense matters even more heading into a season where governance proposals are dominating shareholder activity.

Why Do These Amendments Matter for Boards Right Now?

The corporate legal community has broadly welcomed both amendments as a stabilizing measure after several years of expanding litigation risk around conflicted transactions and inspection demands. According to commentary published through the Harvard Law School Forum on Corporate Governance, the changes are intended to decrease costly, repetitive litigation while giving directors greater certainty when navigating routine governance decisions.

Are Directors’ AI Chat Conversations Discoverable in Litigation?

Whether an AI chat conversation is discoverable now depends on who used the AI, why, and under whose direction — three 2026 rulings show courts drawing that line case by case rather than applying a single blanket rule. This is now a live governance question, not a hypothetical one, for any director or officer who uses AI tools to think through a sensitive decision.

In U.S. v. Heppner (S.D.N.Y., February 2026), a defendant’s exchanges with a consumer version of an AI assistant were ruled discoverable because the individual used the tool independently, without any attorney direction that might have triggered privilege. In Fortis Advisors v. Krafton (Delaware Court of Chancery, March 2026), a Delaware court weighed a CEO’s AI chat exchanges about an earnout-avoidance strategy as direct evidence in the dispute. By contrast, in Warner v. Gilbarco Inc. (E.D. Michigan, February 2026), a federal court found that AI use in anticipation of litigation qualified for work-product protection.

What Do These Three Rulings Mean in Practice for Directors Using AI?

The dividing line across all three cases is attorney involvement and litigation anticipation: AI conversations run independently by a non-lawyer executive, with no legal direction and no anticipated dispute, are the most vulnerable to discovery. Conversations initiated at counsel’s direction, or clearly prepared because litigation was already anticipated, have a stronger — though still developing — claim to privilege or work-product protection.

For a director or officer, this means an offhand ChatGPT session used to think through a strategic maneuver, like the earnout question in Fortis v. Krafton, can become a courtroom exhibit with far less friction than a memo drafted with a lawyer’s guidance. Courts are still building doctrine here, and outcomes are varying by jurisdiction, which makes a written internal policy more valuable than waiting for settled case law — a concern that now sits alongside AI and geopolitical risk as top-of-mind issues for corporate boards.

How Should Boards Set AI Usage Policies to Manage Legal Risk?

Boards should adopt a written AI usage policy that treats AI conversations about company matters as potential business records from the moment they are created, not as informal brainstorming exempt from discovery.

• Route sensitive AI use through counsel — when a director or officer wants to use AI to think through a conflicted transaction, litigation risk, or restructuring, looping in counsel first strengthens any later privilege claim.
• Ban consumer AI accounts for board-related matters — require enterprise AI tools with logging and retention controls the company can manage, rather than personal ChatGPT or Claude accounts outside IT’s visibility.
• Update Section 144 approval templates now — confirm with counsel that conflicted-transaction approval processes explicitly track the amended statutory safe harbor language, not the pre-2025 standard.
• Prepare for narrower, sharper Section 220 demands — since broad fishing-expedition-style requests are harder to sustain, expect stockholder demands to be more targeted, and keep records organized accordingly.

How Does This Fit Into the Broader 2026 Governance Trend?

Both developments — the Section 144/220 amendments and the AI-discoverability rulings — reflect the same underlying shift in 2026 corporate governance: legislatures and courts are racing to catch up with how directors and officers actually make decisions today, using new tools and facing new litigation tactics that older statutes and precedents never anticipated.

Delaware’s amendments were a direct legislative response to perceived litigation abuse under the old Section 144 and Section 220 standards. The AI-discoverability cases are the judicial equivalent: courts confronting a genuinely new fact pattern — an executive’s private conversation with a machine — and reaching for existing doctrines on privilege, work product, and independent action because no AI-specific framework yet exists.

For general counsel and corporate secretaries, the practical overlap is significant. Both issues ultimately turn on documentation quality: a conflicted transaction survives scrutiny when the approval process is well-documented under the new Section 144 safe harbor, and an AI-assisted decision survives discovery disputes more cleanly when it was made under counsel’s direction and recorded as such. Boards that treat 2026 as the year to formalize both practices — updated transaction approval templates and a written AI usage policy — will be far better positioned than those waiting for either area of law to fully settle.

Expect both trends to accelerate through the rest of 2026: more states are watching Delaware’s amendments as a template, and more courts will confront AI-discoverability questions as AI tools become a routine part of how officers and directors work through sensitive decisions.

Frequently Asked Questions

What is the Delaware Section 144 safe harbor?

The Section 144 safe harbor, created by 2025’s Senate Bill 21, is a statutory process that protects conflicted-interest transactions from later legal challenge when they are approved by a disinterested, informed board majority or through a fully informed stockholder vote.

Can a stockholder still demand company books and records under amended Section 220?

Yes, but the amendment narrows the scope of what can be demanded, limiting broad pre-litigation fishing expeditions while still allowing access to records reasonably related to a stockholder’s specific, proper purpose.

Is a ChatGPT conversation with a company executive protected by attorney-client privilege?

Not automatically — 2026 rulings show privilege protection depends on whether the conversation was directed by counsel or conducted independently by a non-lawyer executive, with independent, undirected use being far more likely to be ruled discoverable.

Should boards ban the use of AI tools for governance matters?

Outright bans are impractical and rarely enforceable; a better approach is requiring enterprise AI tools with proper logging, routing sensitive use through counsel, and treating every AI conversation about company business as a potential discoverable record.

Son Güncelleme / Last Updated: July 27, 2026. Sources: Harvard Law School Forum on Corporate Governance (July 2026); Delaware Senate Bill 21 (2025); U.S. v. Heppner (S.D.N.Y. 2026); Fortis Advisors v. Krafton (Del. Ch. 2026); Warner v. Gilbarco Inc. (E.D. Mich. 2026).