Hussain v. Mansoor: Connecticut Court Excises AI Hallucinations From Special Motion to Dismiss, Decides Case on Merits | Advocate Prakhar

Case at a Glance

Background

Mohamed Hussain, a nurse practitioner and sole member of VCare Family Practice LLC, and employee Aiholaney Garcia, filed a defamation and invasion of privacy action against Mansoor Quraishi, a former business associate and technology provider. Quraishi had made ten Facebook posts between July 2 and July 11, 2025, making a series of extremely serious and specific allegations: accusing Hussain of identity fraud, theft, defrauding Medicare and Medicaid, transporting controlled substances internationally, using a terminally ill doctor as a fraudulent medical director, and sexually assaulting a patient. The posts were public, explosive, and specific.

Quraishi, who was also an attorney, filed a special motion to dismiss under Connecticut General Statutes § 52-196a (the state’s anti-SLAPP statute), arguing his Facebook posts were protected speech in connection with a matter of public concern. He admitted at the March 2026 hearing that he used AI to write the motion. The court found eight problematic citations in the brief and issued an order to show cause before the hearing. At the hearing, Quraishi chose to rest on the statutory framework rather than defend the AI-generated content.

The AI Issue

The Connecticut court’s approach here was practical and proportionate. Rather than striking the motion entirely or sanctioning Quraishi immediately, it excised the AI-generated portions and decided the motion on the statutory framework alone. This preserved the defendant’s statutory right to file an anti-SLAPP motion while preventing the court from being misled by hallucinated authorities. The court also noted that the plaintiff did not object to this approach. The motion was ultimately denied on the merits — the Facebook posts were not protected speech because the court found probable cause that the defamation and privacy claims would succeed.

What the Court Decided

• AI-generated portions of the motion to dismiss excised; case decided on statutory framework under Conn. Gen. Stat. § 52-196a.
• Motion to dismiss denied: plaintiffs demonstrated probable cause that they would prevail on defamation, intentional infliction of emotional distress, and invasion of privacy claims.
• Defamation: the Facebook posts were false statements of fact presented as true to a general audience; not protected opinion; probability of truth defence insufficient at this stage.
• Invasion of privacy: the posts attributed highly offensive and false criminal conduct to the plaintiffs in a manner that a reasonable person would find deeply offensive.
• Attorney fees sought by plaintiffs denied: the court could not find the anti-SLAPP motion was “frivolous and solely intended to cause delay” under the applicable standard.
• Court did not impose monetary sanctions on Quraishi at this stage despite the AI-citation issues.

“At the hearing, Attorney Quraishi admitted that he used AI to write the special motion to dismiss.”

— Graff, J., Superior Court of Connecticut, March 31, 2026

The India Angle

Indian Law Equivalent

India does not have a general anti-SLAPP statute comparable to Connecticut’s § 52-196a, though Section 499 of the IPC (defamation) and Section 503 (criminal intimidation) provide some protections for speech, and courts have developed doctrines of fair comment and qualified privilege. A defendant who uses AI to draft a defamation defence in India would face the same professional obligations as in this case — and the same risk of hallucinated citations undermining an otherwise legitimate legal argument. Defamation claims in India proceed in civil courts under the Tort of Defamation (common law) and criminal courts under Section 499 IPC; the pleading obligations in both are demanding and require accurate citation of applicable principles.

Bar Council Rules

The BCI Rules require advocates to act with professional diligence and accuracy. An advocate who openly admits to a court that AI wrote their motion, and that they submitted it without checking the eight problematic citations it contained, is in a difficult professional position. The court’s decision not to sanction Quraishi financially in this instance — partly because he admitted and rested on the statutory framework — reflects the leniency courts extend to advocates who are transparent about AI use. But the underlying professional obligation was clearly violated: the motion contained eight citation defects that should have been caught before filing.

Practical Advice for Indian Advocates

• When AI generates a legal motion with citation problems that are pointed out by the court, the best strategy is transparency: admit the AI use, offer to proceed on the statutory or factual framework without relying on the defective citations, and avoid appearing defensive — as Quraishi’s approach in this case demonstrated, transparency preserves the statutory right while avoiding the worst sanction outcomes.
• Anti-SLAPP and defamation defence motions require precise analysis of the specific statements made and the applicable privilege categories — AI-generated arguments in this area tend to be generic and may mischaracterise the specific privileges available under the applicable state or national law.
• Explicit admission of AI use to a court, while it does not cure the professional obligation to verify citations, demonstrates candour that courts often credit in sanction analyses — concealment of AI use is treated as more serious misconduct than transparent acknowledgment.

Quick Takeaways

• Courts have the option of excising AI-generated problematic portions and deciding motions on the remaining statutory or factual framework — this preserves procedural rights while refusing to be misled by hallucinations.
• Transparent admission of AI use, combined with willingness to proceed on the statutory framework, mitigates sanction risk even when citation errors are substantial.
• Facebook posts making specific criminal accusations against identifiable individuals are highly likely to fail anti-SLAPP protection — they are not speech on a matter of public concern but targeted personal attacks on private individuals.

Deep Dive: The Excision Remedy — A Proportionate Response to AI Citation Failures

The Hussain v. Mansoor case introduces a remedy for AI-hallucination misconduct that has not appeared prominently in other reported cases: excision of the AI-generated portions, followed by adjudication on the remaining valid framework. This approach is significant because it avoids two extremes — wholesale striking of the motion (which would punish the client for counsel’s error) and accepting the AI-generated content (which would allow fabricated citations to influence the outcome).

The excision remedy works when there is a core statutory or factual framework that is independent of the AI-generated citations. Here, the anti-SLAPP statute has a specific analytical framework that does not depend on case law to function — the court can apply the statute to the facts without needing the cases that Quraishi’s AI hallucinated. When the movant agrees to rest on that framework, the court can proceed to a substantive decision that is based entirely on authenticated material.

This remedy is particularly appropriate where: (1) the movant admits AI use transparently; (2) the statutory or regulatory framework is clear and self-applying; (3) the AI-hallucinated citations were supporting arguments rather than essential elements; and (4) the opposing party does not object to proceeding. Where any of these conditions is absent — particularly where the movant is defensive about AI use or where the hallucinated citations were central to the legal argument — excision may not be a viable option and the court may need to strike the filing entirely.

For Indian advocates, the excision remedy has a practical application: when you discover that a brief you filed contains AI-hallucinated citations, consider filing a notice of errata or correction that specifically identifies the defective citations, withdraws reliance on them, and clarifies that the remaining arguments rest on the statutory or regulatory framework and the verified citations. This is the functional equivalent of the excision the Connecticut court performed, and it demonstrates the same transparency and professional responsibility that the court credited in this case.