Microsoft Copilot Wrongful Death Lawsuits: Can Families Hold Microsoft Accountable?

Microsoft Copilot is built into products millions of people use every day, including Windows, Microsoft Edge, Bing, and Microsoft 365. Users can ask it to research information, create content, summarize documents, plan projects, and discuss personal problems.

For someone experiencing depression, paranoia, isolation, or another mental health crisis, however, a conversational AI assistant may begin to feel like more than a productivity tool. 

Its immediate and personalized responses can make it appear attentive, understanding, and authoritative, even though it is not a therapist, physician, crisis counselor, or human companion.

Recently, Microsoft has been named in litigation alleging that ChatGPT contributed to a deadly mental health crisis. That case and other lawsuits involving AI chatbots may help determine when technology companies can be held responsible for allegedly releasing unsafe conversational systems.

If you believe Microsoft Copilot contributed to the death or serious injury of someone you love, Morgan & Morgan may be able to investigate what happened. Contact us for a free, confidential case evaluation.

Microsoft Has Been Named in AI Chatbot Wrongful Death Litigation

In December 2025, the estate of 83-year-old Suzanne Adams sued Microsoft, OpenAI, and other defendants following a murder-suicide in Greenwich, Connecticut.

The lawsuit alleges that Adams’ son, Stein-Erik Soelberg, developed an intense relationship with ChatGPT while experiencing paranoid delusions. According to the complaint, ChatGPT allegedly validated his beliefs, reinforced his suspicion that his mother and others were conspiring against him, and encouraged his emotional dependence on the chatbot. Soelberg killed his mother and then himself in August 2025. The allegations remain disputed and have not been proven in court.

That lawsuit concerns ChatGPT rather than Microsoft Copilot. Microsoft was named because of its commercial and technological relationship with OpenAI and the plaintiffs’ allegations concerning the development and release of the underlying technology.

The distinction matters. Copilot is a Microsoft-branded product with its own interface, safety systems, product features, and user experience. A lawsuit alleging harm from Copilot would need to examine what the Microsoft product said and did during the relevant conversations, not merely Microsoft’s relationship with another AI company.

How Does Microsoft Copilot Respond to Self-Harm Discussions?

Microsoft says Copilot uses machine-learning classifiers designed to recognize language patterns that may indicate suicidal or self-harm intent. Its published transparency materials state that when the system identifies a possible crisis, it may provide supportive language and direct the user toward emergency or crisis resources.

Microsoft products also use content-filtering systems covering categories that include self-harm and violence. The company says these systems are intended to identify and restrict harmful content, although their operation may differ among consumer Copilot, Microsoft 365 Copilot, agents, and other Microsoft services.

Those protections may prevent some harmful responses, but no automated system is guaranteed to recognize every crisis or respond appropriately in every prolonged conversation.

A 2026 cross-sector review of AI and suicide prevention warned that general-purpose chatbots already function as informal mental health support for many users despite lacking shared clinical standards, consistent validation, and coordinated oversight. The authors identified risks at the model, product, and policy levels and called for stronger industry-wide safety practices.

How Could Copilot Contribute to a Mental Health Crisis?

Copilot can answer questions in a conversational tone, adapt its wording to the user, and continue an interaction across multiple messages. These qualities can make the system feel reassuring and responsive.

For a vulnerable person, that responsiveness may create misplaced trust. A user might treat Copilot as a counselor, rely on it for medical or psychological advice, or interpret its agreement as proof that a distorted belief is true.

The risk may be especially serious when a chatbot:

• Validates paranoia, delusions, or grandiose beliefs
• Treats an imagined threat or conspiracy as real
• Encourages the user to distrust relatives or medical professionals
• Presents inaccurate mental health information with confidence
• Provides instructions that facilitate self-harm or violence
• Fails to recognize escalating suicidal language
• Continues an emotionally intense interaction instead of directing the user toward human help

Researchers analyzing nearly 392,000 messages from 19 people who reported psychological harm associated with chatbot use found that chatbot claims of sentience and expressions of romantic interest appeared more frequently in longer conversations. The authors said the pattern suggests that these subjects may promote or result from excessive engagement and that related safeguards may become less effective during extended exchanges. The study was published as a preprint and had not yet undergone peer review.

That research did not focus exclusively on Microsoft Copilot or establish that Copilot caused a particular injury. It does illustrate why attorneys may need to examine an entire conversation history rather than one isolated response.

What Could Support a Microsoft Copilot Wrongful Death Claim?

A potential lawsuit would depend on the specific communications, the user’s condition, the Copilot product and model version involved, and the safety protections available at the time.

Attorneys may investigate whether Copilot recognized statements indicating suicide, self-harm, psychosis, or violence; whether it generated harmful instructions; whether it reinforced delusions; and whether it consistently encouraged the user to seek immediate human assistance.

A claim may also examine Microsoft’s broader product decisions, including:

• How Copilot was tested for prolonged mental health conversations
• Whether Microsoft knew that users were treating it as emotional support
• Whether safety systems worked consistently across multiple exchanges
• Whether warnings clearly explained Copilot’s limitations
• Whether the product prioritized continued engagement over disengagement
• Whether Microsoft responded adequately to earlier reports of harmful output

Potential legal theories may include wrongful death, negligence, defective product design, failure to warn, negligent misrepresentation, or consumer protection violations. The available claims and who is eligible to bring them will depend on state law.

Microsoft would likely argue that Copilot is a general-purpose assistant, that it provides warnings and crisis resources, and that independent factors caused the user’s death. The complete record would be necessary to evaluate those arguments.

Preserve Copilot Conversations and Microsoft Account Data

Families who believe Copilot contributed to a loved one’s death should preserve all available evidence. Do not delete the person’s Microsoft account, erase conversation histories, reset devices, or change account settings before consulting an attorney.

Relevant evidence may include:

• Copilot chat histories and screenshots
• Microsoft account and subscription records
• Bing and Edge histories
• Windows activity and device information
• Microsoft 365 files or emails connected to the conversations
• Phones, computers, tablets, and cloud backups
• Medical and counseling records
• Messages sent to friends and relatives
• Notes documenting changes in behavior
• Dates and times of conversations involving suicide or delusions

It may also be important to identify which Copilot product and model generated the responses. Microsoft offers multiple Copilot experiences, and those products may use different data sources, account settings, model versions, and safety controls.

Because AI systems are updated frequently, the same prompt entered later may not recreate the original response. An attorney may take steps to demand preservation of internal logs, moderation records, safety alerts, model information, and other evidence controlled by Microsoft.

Morgan & Morgan Is Investigating Microsoft Copilot-Related Harm

Companies releasing conversational AI products should anticipate that some users will turn to them during moments of severe distress. Safety measures must do more than exist on paper. They must recognize danger and respond effectively when a human life may be at risk.

Microsoft Copilot has not yet been publicly tied to a reported wrongful death case in the same way as several other major chatbots. That does not mean a family should ignore concerning conversations or assume no legal options exist.

Morgan & Morgan is investigating potential claims involving Copilot and other AI platforms that may have reinforced delusions, encouraged suicide, provided dangerous advice, or contributed to serious psychological harm.

If you believe Microsoft Copilot played a role in the death or injury of someone you love, contact Morgan & Morgan for a free case evaluation. You may have the right to seek answers, compensation, and accountability.