OpenAI’s ChatGPT will soon allow ‘erotica’ for adults in major policy shift
Sam Altman, chief executive officer of OpenAI Inc., during a media tour of the Stargate AI data center in Abilene, Texas, US, on Tuesday, Sept. 23, 2025.
Kyle Grillot | Bloomberg | Getty Images
Adult ChatGPT users can soon access a less censored version of the artificial intelligence chatbot, which will include erotic materials, OpenAI CEO Sam Altman has announced in an apparent policy shift.
“In December, as we roll out age-gating more fully and as part of our ‘treat adult users like adults’ principle, we will allow even more, like erotica for verified adults,” Altman said in a social media post on Tuesday.
Though it remains unclear what material will qualify as permitted erotica, the move could represent a major shift in OpenAI’s policy, which formerly prohibited such content in most contexts.
According to Altman, existing versions of ChatGPT were made “pretty restrictive” to protect users from mental health risks, but that approach made the chatbot “less useful [and enjoyable to many users who had no mental health problems.
“Now that we have been able to mitigate the serious mental health issues and have new tools, we are going to be able to safely relax the restrictions in most cases,” he said.
Those “new tools” appear to refer to safety features and parental controls rolled out last month to address concerns over how the chatbot was impacting young users’ mental health.
However, as safeguards for minors expand, it appears that Altman is ready for ChatGPT to take a looser approach for adults.
OpenAI hinted at a shift in February when language on its “Model Spec” page was updated to clarify that, in order to “maximize freedom” for users, only sexual content involving minors was prohibited. Still, erotica was considered to be “sensitive content” to be generated only in certain permitted contexts.
Besides the rollout in December, Altman also said a new version of ChatGPT will launch in the coming weeks, allowing the chatbot to adopt more distinct personalities — building on updates in the latest GPT‑4o version.
“If you want your ChatGPT to respond in a very human-like way, or use a ton of emoji, or act like a friend, ChatGPT should do it,” he said. “But only if you want it.“
Growth vs. safety
After Altman’s post on Tuesday, some social media users were quick to point out his previous statements suggesting that ChatGPT wouldn’t implement sexualized chat features, unlike rival models such as xAI’s Grok.
In an August interview, independent tech journalist Cleo Abram asked Altman to give an example of a decision he had made that was best for the world, but not for winning the AI race.
“Well, we haven’t put a sex bot avatar in ChatGPT yet,” Altman said in an apparent nod to provocative AI companions released by Elon Musk’s xAI.
Altman’s policy shift comes at a notable time for OpenAI, as it already faces increased scrutiny over its safety practices. In September, the Federal Trade Commission launched an inquiry into several tech companies, including OpenAI, over potential risks to children and teenagers.
That followed a lawsuit from a California couple who alleged that ChatGPT contributed to their 16-year-old son’s suicide.
OpenAI on Tuesday also announced an eight-member expert council on well-being and AI to advise the company on how artificial intelligence affects users’ mental health, emotions and motivation.
The council will guide OpenAI in defining what healthy AI interactions look like through check-ins and recurring meetings, the company said.
Beta Technologies shares surged more than 9% after air taxi maker Eve Air Mobility announced an up to $1 billion deal to buy motors from the Vermont-based company.
Eve, which was started by Brazilian airplane maker Embraer and is now under Eve Holding, said the manufacturing deal could equal as much as $1 billion over 10 years. The Florida-based company said it has a backlog of 2,800 vehicles.
Shares of Eve Holding gained 14%.
Eve CEO Johann Bordais called the deal a “pivotal milestone” in the advancement of the company’s electric vertical takeoff and landing, or eVTOL, technology.
“Their electric motor technology will play a critical role in powering our aircraft during cruise, supporting the maturity of our propulsion architecture as we progress toward entry into service,” he said in a release.
Mateusz Slodkowski | SOPA Images | Lightrocket | Getty Images
Amazon’s cloud unit on Tuesday announced AI-enabled software designed to help clients better understand and recover from outages.
DevOps Agent, as the artificial intelligence tool from Amazon Web Services is called, predicts the cause of technical hiccups using input from third-party tools such as Datadog and Dynatrace. AWS said customers can sign up to use the tool Tuesday in a preview, before Amazon starts charging for the service.
The AI outage tool from AWS is intended to help companies more quickly figure out what caused an outage and implement fixes, Swami Sivasubramanian, vice president of agentic AI at AWS, told CNBC. It’s what site reliability engineers, or SREs, do at many companies that provide online services.
SREs try to prevent downtime and jump into action during live incidents. Startups such as Resolve and Traversal have started marketing AI assistants for these experts. Microsoft’s Azure cloud group introduced an SRE Agent in May.
Rather than waiting for on-call staff members to figure out what happened, the AWS DevOps Agent automatically assigns work to agents that look into different hypotheses, Sivasubramanian said.
“By the time the on-call ops team member dials in, they have an incident report with preliminary investigation of what could be the likely outcome, and then suggest what could be the remediation as well,” Sivasubramanian told CNBC ahead of AWS’ Reinvent conference in Las Vegas this week.
Commonwealth Bank of Australia has tested the AWS DevOps Agent. In under 15 minutes, the software found the root cause of an issue that would have taken a veteran engineer hours, AWS said in a statement.
The tool relies on Amazon’s in-house AI models and those from other providers, a spokesperson said.
AWS has been selling software in addition to raw infrastructure for many years. Amazon was early to start renting out server space and storage to developers since the mid-2000s, and technology companies such as Google, Microsoft and Oracle have followed.
Since the launch of ChatGPT in 2022, these cloud infrastructure providers have been trying to demonstrate how generative AI models, which are often training in large cloud computing data centers, can speed up work for software developers.
Over the summer, Amazon announced Kiro, a so-called vibe coding tool that produces and modifies source code based on user text prompts. In November, Google debuted similar software for individual software developers called Antigravity, and Microsoft sells subscriptions to GitHub Copilot.
WATCH: Amazon rolls out AI-powered tools to help big AWS customers update old software
-
Sports2 years agoStory injured on diving stop, exits Red Sox game
-
Sports3 years ago‘Storybook stuff’: Inside the night Bryce Harper sent the Phillies to the World Series
-
Sports2 years agoGame 1 of WS least-watched in recorded history
-
Sports3 years agoButton battles heat exhaustion in NASCAR debut
-
Sports3 years agoMLB Rank 2023: Ranking baseball’s top 100 players
-
Sports4 years ago
Team Europe easily wins 4th straight Laver Cup
-
Environment3 years agoJapan and South Korea have a lot at stake in a free and open South China Sea
-
Environment1 year agoHere are the best electric bikes you can buy at every price level in October 2024