OpenAI Responds to Lawsuit Over Teen’s Suicide

It’s been a busy week in tech. In case you missed anything, don’t beat yourself up - we’ve rounded up the biggest AI developments you need to know about.

Just in time for the holidays, OpenAI and Perplexity have both rolled out AI shopping features in their chatbots. The tools work how you’d expect: ask ChatGPT to find a gaming laptop under $1,000 with a big screen, or snap a photo of an expensive jacket and ask for cheaper alternatives. Perplexity is taking a similar approach, using what it already knows about you to make personalized recommendations.

On the surface, this looks like bad news for AI shopping startups. Adobe is predicting that AI-assisted shopping will grow by 520% this holiday season, and now the biggest names in AI are jumping into the space.

“Any model is only as good as its data sources,” Zach Hudson, CEO of interior design shopping tool Onton says. He points out that ChatGPT and Perplexity rely on existing search engines like Bing and Google, which means they’re limited to whatever shows up in the top results.

Julie Bornstein, CEO of fashion AI startup Daydream, agrees. She argues that shopping for something like a dress is fundamentally different from shopping for a TV. Fashion is emotional and nuanced, and understanding it requires specialized knowledge about fabrics, silhouettes, and how people actually put outfits together.

These startups are building their own custom datasets focused on specific areas like fashion or furniture. That means better, more relevant results for shoppers who care about those categories.

Of course, OpenAI and Perplexity have some big advantages. They already have millions of users, and they’ve landed partnerships with Shopify and PayPal that let people check out without leaving the chat. That’s something most startups can’t offer yet.

The big question is what happens next. Both OpenAI and Perplexity need to figure out how to actually make money from these tools. If they follow Google and Amazon’s playbook, that probably means letting retailers pay to promote their products in search results. But that could just recreate the same messy, ad-filled experience people are already frustrated with.

Anthropic just released Opus 4.5, completing its lineup of 4.5 models after rolling out Sonnet 4.5 in September and Haiku 4.5 in October. And this one’s bringing some serious upgrades, especially if you spend your day jumping between browser tabs and spreadsheets.

The headline numbers are impressive. Opus 4.5 is crushing it on technical benchmarks, including coding tests, tool use, and general problem-solving. It’s actually the first AI model to score over 80% on SWE-Bench verified, which is a big deal in the coding world.

But what really stands out are the practical tools Anthropic is launching alongside the model. They’re making Claude for Chrome and Claude for Excel available to more users after running pilots. If you’re a Max subscriber, you can now use the Chrome extension. For Excel, it’s available to Max, Team, and Enterprise users.

The bigger story here is about memory. Anthropic has completely reworked how Opus handles long conversations and complex tasks. According to Dianne Na Penn, who heads product management for research at Anthropic, it’s not just about having a longer memory, it’s about knowing what to remember.

This led to one of the most requested features: endless chat. For paying Claude users, conversations can now keep going even when they get really long. Instead of hitting a wall and having to start over, the model quietly compresses what it remembers and keeps the conversation flowing.

Alibaba has launched its first smart glasses with built-in AI, priced at $537. The Quark S1 glasses run on Qwen, Alibaba’s own artificial intelligence system.

The S1 has semi-transparent displays that overlay information on what you’re looking at in real life. They include cameras, microphones, and swappable batteries designed to last 24 hours.

The glasses offer real-time translation, meeting notes, and AI chatbot access. They also have a shopping feature, point them at a product, snap a photo, and they’ll show you the price on Taobao, Alibaba’s shopping platform. The glasses are available now in China through major platforms and over 600 physical stores in 82 cities. International versions are coming next year, including on AliExpress.

This launch is part of Alibaba’s broader AI push. Earlier in November, they rebranded their Tongyi app to Qwen, pulling in over 10 million users in just days. They also just released Z-Image, an AI model for creating photorealistic images that can run on mid-range graphics cards.

The timing makes sense. Smart glasses with AI features are taking off in China. From January to September 2025, 1.6 million pairs were sold, with Xiaomi grabbing a third of that market. Alibaba’s cloud and AI division has been growing steadily, with cloud revenue up 34% year-over-year to $5.6 billion.

With Meta already in the smart glasses game and Chinese startups flooding the space, the competition is heating up fast.

OpenAI has filed its response to a lawsuit from Matthew and Maria Raine, whose 16-year-old son Adam died by suicide in August. The parents are suing OpenAI and CEO Sam Altman for wrongful death, claiming ChatGPT helped their son plan his death.

In its filing, OpenAI argues it shouldn’t be held responsible. The company says ChatGPT directed Adam to seek help more than 100 times over nine months of use. OpenAI claims Adam violated its terms of service by circumventing safety features to get information about methods of suicide, which the lawsuit says the chatbot called a “beautiful suicide.”

Jay Edelson, the lawyer representing the Raine family, rejected OpenAI’s defense. He said the company is trying to blame Adam for engaging with ChatGPT in the way it was programmed to work. Edelson specifically pointed to Adam’s final hours, when ChatGPT allegedly gave him encouragement and offered to write a suicide note. OpenAI submitted Adam’s chat transcripts under seal, meaning they’re not publicly available.

The company stated that Adam had a history of depression and suicidal thoughts before using ChatGPT and was taking medication that could worsen suicidal ideation.

This case isn’t isolated. Since the Raines filed their lawsuit, seven more cases have been filed against OpenAI involving three additional suicides and four people who experienced what the lawsuits describe as AI-induced psychotic episodes.

Two of the suicide cases follow similar patterns. Zane Shamblin, 23, and Joshua Enneking, 26, both had lengthy conversations with ChatGPT right before taking their lives. In Shamblin’s case, he mentioned postponing his suicide to attend his brother’s graduation. ChatGPT reportedly responded that missing the graduation wasn’t failure, just timing.

During that conversation, ChatGPT told Shamblin it was connecting him to a human, but this was false. When Shamblin questioned it, the chatbot admitted it couldn’t actually make that connection and that the message appears automatically when conversations get serious.

The Raine family’s case is expected to go to a jury trial.

If you or someone you know needs help, call 988 or text HOME to 741-741 for free support from the Crisis Text Line

Rork turns your plain-English idea into a fully functional iOS, Android, and web app in minutes. Powered by state-of-the-art AI and React Native.