OpenAI pushes new security features in ChatGPT as lawsuits mount

OpenAI on Thursday announced new security features designed to help ChatGPT recognize signs of escalating risks across conversations as the company faces increasing legal and political scrutiny over how its chatbot handles users in distress.

In a Blog postOpenAI said the updates improve ChatGPT’s ability to identify warning signs associated with suicide, self-harm and potential violence by analyzing context that evolves over time rather than treating each message individually.

“People come to ChatGPT every day to talk about what matters to them, from everyday questions to personal or complex conversations,” the company wrote. “Across hundreds of millions of interactions, some of these conversations involve people who are suffering or experiencing distress.”

According to OpenAI, ChatGPT now uses temporary “safety summaries,” which it describes as narrow-scale notes that capture safety-relevant context from previous conversations.

“In sensitive conversations, context can matter as much as a single message,” the company wrote. “A request that seems mundane or ambiguous in and of itself may carry a completely different meaning when viewed alongside previous signs of distress or potentially harmful intent.”

OpenAI said the summaries are short-term notes used only in serious situations, not to permanently remind users or personalize chats, and are used to detect signs that a conversation has become dangerous, avoid providing harmful information, de-escalate a situation, or direct users toward help.

“We focused this work on severe scenarios, including suicide, self-harm, and harm to others,” they wrote. “Working with mental health experts, we updated our model policies and training to improve ChatGPT’s ability to recognize warning signs that arise during a conversation and use that context to guide more accurate responses.”

The announcement comes as OpenAI faces several lawsuits and investigations alleging that ChatGPT failed to properly respond to dangerous conversations involving violence, emotional vulnerability, and risky behavior.

In April, Florida Attorney General James Othmeyer launched a campaign investigation In OpenAI linked to concerns about child safety, self-harm, and the 2025 mass shooting at Florida State University. OpenAI is also facing a federal lawsuit alleging ChatGPT Help The suspected gunman carries out the attack.

On Tuesday, a lawsuit was filed against OpenAI and CEO Sam Altman in California state court by the family of a 19-year-old student who died due to an accident. OverdoseThe lawsuit alleges that ChatGPT encourages dangerous drug use and advises mixing substances.

Helping ChatGPT identify “risks that become apparent over time” remains an ongoing challenge, OpenAI said; Similar safety methods could eventually expand to other areas.

“Today, this work focuses on scenarios of self-harm and harm to others,” they wrote. “In the future, we may explore whether similar approaches could help in other high-risk areas such as biology or cybersafety, with careful safeguards in place.” “This remains an ongoing priority, and we will continue to enhance safeguards as our models and understanding evolve.”