ChatGPT chats were showing up on Google, but OpenAI says its all good now
From understanding astrophysics to asking for advice to improve mental health, conversations with ChatGPT were showing up on Google if you added the term 'site: chatgpt.com/share' to your search query. While most of these conversations were uninteresting and simple, some of them were very personal. For example, several chats which discussed issues like mental health, sex life, career advice, addiction, physical abuse and other serious topics could be brought up with a simple Google search.
While ChatGPT does not make conversations public by default, chats that were shared using the 'Share' button in the app and website were indexed by search engines like Google. For those unaware, ChatGPT's share feature creates a link, which you can send to others over WhatsApp, Instagram, or Facebook so they can view your conversation with the AI chatbot.
The feature also allows people who have the link to continue the conversation with ChatGPT if they want to. One thing to note is that even if you delete the link, it might still be visible on cached pages.
First discovered by Fast Company, the publication says that around 4,500 conversations were visible on Google site search, but most of them did not contain personally identifiable information. However, if someone included a name or phrase in a chat, it may show up in search results.
The publication goes on to say that the number of leaked chats is probably much higher than the stated figure, as Google may not index all conversations. In a statement to PCMag, a Google spokesperson said that OpenAI is responsible for making these chats visible to search engines.
We just removed a feature from @ChatGPTapp that allowed users to make their conversations discoverable by search engines, such as Google. This was a short-lived experiment to help people discover useful conversations. This feature required users to opt-in, first by picking a chat… pic.twitter.com/mGI3lF05Ua
— DANΞ (@cryps1s) July 31, 2025
In a post on X, an OpenAI employee said that they have now removed the feature that allowed users to make their conversations discoverable by search engines. He went on to say that this 'was a short-lived experiment' and that it 'introduced too many opportunities for folks to accidentally share things they didn't intend to.'
OpenAI's Shared Links FAQ states that ChatGPT chats are not made public unless you tick the 'Make this chat discoverable' option when using the AI chatbot's built-in share feature. As it turns out, many ChatGPT users may have accidentally ticked this box, thinking it necessary to do so if they want their conversation to be visible to others.
In case you are wondering, these shared links can be deleted at any time. To do so, head over to ChatGPT settings and, under the 'Data controls' section, click the 'Manage' button that appears to the right of the 'Shared links' option. Here, you will see how many chats you have shared publicly and the option to delete these links.
OpenAI CEO Sam Altman had also said that since there is currently no legal or policy framework for AI, users should not expect any legal confidentiality for their conversations with ChatGPT. This means that none of your chats are actually private and OpenAI might be forced to share your conversation with the AI chatbot in court if asked to do so.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
India Today
2 minutes ago
- India Today
Google takes a dig at Apple's AI Siri delays ahead of Pixel 10 launch, says change your phone now
Google is using Apple's delay in launching its new AI-powered Siri as a chance to push its upcoming Pixel 10 series. In a new ad, the company throws subtle shade at Apple for not delivering on its promise of AI features, which were showcased over a year ago. The ad, though never naming Apple directly, clearly refers to it, saying: 'If you buy a new phone because of a feature that's coming soon, but it's been coming soon for a full year, you could change your definition of soon. Or you could just change your phone.' This playful jab comes after Apple confirmed that its personalised Siri — powered by Apple Intelligence — will now arrive in 2026 instead of the originally promised 2025. advertisementOf course, the delay has sparked frustration among Apple users, even leading to lawsuits over misleading marketing. Google, on the other hand, is all set to launch its next-gen Pixel smartphones later this Pixel 10 series is expected to be announced on August 20, with pre-orders opening the same day. Shipping is likely to begin on August 28. Google is expected to launch four models: the Pixel 10, Pixel 10 Pro, Pixel 10 Pro XL, and the Pixel 10 Pro Fold. In addition to the phones, the company could also introduce the Pixel Watch 4 and next-gen Pixel Buds Pixel 10 series: What to expect? For the regular Pixel 10, changes come in both design and camera setup. While the basic shape of the phone remains familiar, the colours are getting a refresh. As per rumours, Google is likely dropping the Peony, Wintergreen, and Porcelain shades from the Pixel 9 series. Instead, users will get new colours: Indigo, Frost, and Limoncello, with Obsidian being the only shade to carry over. Frost is a soft blue, Indigo offers a deep, rich tone, and Limoncello is a bold, almost neon back of the Pixel 10 sees the biggest change. For the first time, a base Pixel model is said to include a triple-camera setup, featuring a 48-megapixel primary sensor, a 12-megapixel ultra-wide, and a 10.8-megapixel telephoto lens. This was previously limited to Pro models. The camera bar design stays mostly the same, though the glass covering it is now those wanting more power, the Pixel 10 Pro and Pixel 10 Pro XL are expected to be powered by Google's latest Tensor G5 chip, built on TSMC's 3nm process. Both phones are rumoured to keep the same display sizes as last year — 6.3 inches for the Pro and 6.8 inches for the Pro XL. Endurance is expected to get a boost too, with the Pixel 10 Pro XL said to pack a 5,200mAh battery, the largest ever in a Pixel Pixel 10 Pro Fold is the fourth phone in the lineup, but naturally, expected to be the most expensive one. It keeps the same design language as the previous Pixel 9 Pro Fold, but includes a few upgrades. The cover screen is rumoured to grow slightly to 6.4 inches, the bezels shrink, and a larger 5,015mAh battery is expected. Charging speeds are getting a small bump too — 23W wired and 15W wireless via the Pixel 10 series phones are expected to run on Android 16 and could feature more built-in AI tools, updated Material 3 visuals, and possibly new widgets. Overall, while the phones don't look wildly different from the Pixel 9 generation, the updates under the hood — including performance, battery, and camera — point to solid improvements.- Ends
Indian Express
2 minutes ago
- Indian Express
OpenAI's Sam Altman teases GPT-5 as ChatGPT platform nears 700 million active users
After months of wait, GPT-5 may have made its first public appearance. On Sunday, August 3, OpenAI CEO Sam Altman posted a screenshot on X showing an interaction with the highly anticipated large language model (LLM). The post showed Altman asking GPT-5, 'What is the most thought-provoking show about AI?' to which the model responded with high praise for the science fiction series 'Pantheon'. turns out yes! — Sam Altman (@sama) August 3, 2025 According to GPT-5, Pantheon is 'cerebral, emotional, and philosophically intense' and holds a '100% critic rating on Rotten Tomatoes.' The animated series does have a 100 per cent critic rating on Rotten Tomatoes. Similar words appear in the reviews of the show. One review calls it 'gripping, cerebral, remarkably high-concept.' GPT-5 is said to have a larger context window than its predecessors. As a result, it is capable of handling more agentic tasks and comes with multimodal capabilities. GPT-5 is also widely rumoured to be adept at coding. The upcoming launch of GPT-5 comes at a time when OpenAI is facing significant pressure from its competitors who have launched new iterations of AI models over the past few months such as Google's Gemini 2.5 Pro , XAI's Grok 4, and Anthropic's Claude Opus 4. However, ChatGPT continues to witness substantial growth with the AI chatbot platform on track to reach 700 million users this week, according to Nick Turley, a vice president at OpenAI and head of the ChatGPT app. This week, ChatGPT is on track to reach 700M weekly active users — up from 500M at the end of March and 4× since last year. Every day, people and teams are learning, creating, and solving harder problems. Big week ahead. Grateful to the team for making ChatGPT more useful and… — Nick Turley (@nickaturley) August 4, 2025 'Every day, people and teams are learning, creating, and solving harder problems. Big week ahead. Grateful to the team for making ChatGPT more useful and delivering on our mission so everyone can benefit from AI,' Turley said in a post on X. Its popularity surged in March this year after OpenAI integrated an image generation feature into ChatGPT that kicked off the viral Ghibli trend on social media. OpenAI chief operating officer Brad Lightcap stated that more than 130 million users generated over 700 million images after launching the image generation feature. Its active user base grew to 500 million that month. Furthermore, the number of ChatGPT subscribers has also increased. According to Lightcap, paying corporate customers on ChatGPT rose from 3 million in June to 5 million last week. ChatGPT users utilise the app more than 12 days each month on average, ranking second only to Google and X, according to data from SensorTower. It also found that users spent an average of 16 minutes per day on the AI chatbot app in the first half of 2025.
&w=3840&q=100)
Business Standard
2 minutes ago
- Business Standard
OpenAI optimises ChatGPT with mental health-focused features: What is new
OpenAI is rolling out a series of mental health-oriented features to its AI chatbot ChatGPT. The company said that these additions are designed to help the AI chatbot better recognise signs of emotional distress and respond more thoughtfully. ChatGPT will also now prompt users to take breaks during long interactions. In a blog post outlining the changes, OpenAI noted that these updates follow the rollback of a recent change that made ChatGPT responses overly agreeable — often prioritising what sounded pleasant over what was genuinely useful. The company said it is now refining how it incorporates feedback and measuring responses based on real-world usefulness, not just user satisfaction at the moment. ChatGPT's mental health-focused update: What is new Emotional distress detection: OpenAI said that it is working on improving its models to better recognise when users might be experiencing emotional or mental distress. In such cases, ChatGPT will aim to respond more appropriately and guide users to reliable, evidence-based resources. The company added that it has partnered with over 90 physicians — including psychiatrists, paediatricians, and general practitioners from more than 30 countries — to create evaluation rubrics for complex, multi-turn conversations. Break reminders: The company has announced that starting today, ChatGPT users will see gentle nudges during longer sessions encouraging them to take a break. OpenAI says it will continue adjusting the timing and tone of these reminders to make them feel more organic. Response to personal questions: In the blog, OpenAI said that when asked sensitive or deeply personal questions such as 'Should I break up with my partner?', ChatGPT will now avoid offering binary answers. Instead, it will guide users through a process of reflection — posing questions and helping them weigh different perspectives. This feature is set to roll out soon.
