# OpenAI Rolls Out Incognito Mode on ChatGPT: Giving Users More Control Over Privacy - Canonical: https://33rdsquare.com/openai-rolls-out-incognito-mode-on-chatgpt-giving-users-more-control-over-privacy/ - Published: 2024-09-03 - Author: Jordan Brown - Categories: [Artificial Intelligence & Machine Learning & ChatGPT](https://33rdsquare.com/category/tech/ai/) --- ChatGPT, the AI chatbot from OpenAI that can generate remarkably human-like conversational text, has exploded in popularity since its launch in November 2022. Millions of fascinated users have been chatting with the bot about everything under the sun. However, all those conversations come at a privacy cost. By default, ChatGPT saves chat logs to train its machine learning models. In response to growing concerns, OpenAI has now introduced an "Incognito Mode" for more private chatting, along with additional transparency controls. ## What Exactly Does the New Incognito Mode Do? When Incognito Mode is enabled in settings, ChatGPT will no longer associate conversations with the user‘s account or save chat history. It provides an experience closer to chatting anonymously without a persistent record. OpenAI will still temporarily retain conversations to monitor for any abuse, but will delete the logs after approximately 30 days. This is similar to how private or incognito browsing works in web browsers like Chrome and Firefox. Browsing history is not recorded. For ChatGPT, no chat logs are linked to an account or used for training. ## Why Is ChatGPT Data Usage Sparking Privacy Worries? By default, all user interactions with ChatGPT are retained by OpenAI. The massive amount of conversational data (over 150 billion words and growing daily) is invaluable for training, optimizing, and improving ChatGPT‘s natural language algorithms. The machine learning models look for patterns and relationships in huge datasets to determine how to generate coherent, relevant, and useful responses. More conversational data allows them to handle a wider range of topics and queries flawlessly. However, using personal conversations to train AI systems without explicit consent raises obvious privacy issues. Chat logs can reveal sensitive information about people‘s lives, thoughts, and identities. Some expert commentators argue OpenAI has an ethical responsibility to implement privacy protections like Incognito Mode. ## Europe Led Investigations Over Data Policies Privacy concerns came to a head in January 2023 when Italy‘s data protection agency banned ChatGPT over violations of privacy laws. The agency argued not enough safeguards were in place for the vast amounts of conversational data being collected. France and Spain soon followed with their own investigations into ChatGPT‘s data practices. The scrutiny in Europe appears to have spurred OpenAI into action to launch Incognito Mode quickly. Other countries are keeping a close watch as ChatGPT usage continues rising globally. ## Upcoming Paid ChatGPT Business Version Also Expands Controls In addition to the free consumer version of ChatGPT, OpenAI plans to launch a paid ChatGPT Business subscription in the coming months. It will be tailored for professional use cases. The Business version will provide further transparency and privacy options for enterprise users. By default, OpenAI will be restricted from accessing conversations to train its models, giving companies full control. But the option will exist to voluntarily allow OpenAI to utilize logs anonymously to improve the system‘s capabilities. ## Striking a Balance Between Privacy and AI Progress The changes illustrate OpenAI‘s efforts to balance user privacy with the need for vast amounts of conversational data to develop more capable and useful AI systems. ChatGPT already utilizes some of the most advanced machine learning algorithms developed to date. Its 175 billion parameter model was trained on over 570 gigabytes of text data, including books, websites, and online conversations. This massive foundation of knowledge allows it to draw connections between concepts and synthesize responses like a human. But there are still many limitations, such as bias, factual errors, and lack of deeper reasoning skills. Continued progress requires ongoing learning from new conversations and content. Completely restricting data access would stunt advancement. But ignoring privacy concerns poses dangers as well. **Table 1: Key ChatGPT Data and Usage Statistics** | Statistic | Value | | --- | --- | | Launch Date | November 30, 2022 | | Conversations per day | millions | | Training parameters | 175 billion | | Training data size | 570 GB | | User messages seen | >150 billion words | | Average response time | < 1 second | | Common use cases | Q&A, dialog, explanations | Experts argue a balanced approach that puts users in control through options like Incognito Mode provides the ideal path forward. Transparency, consent, and thoughtful openness will grow even more crucial as AI capabilities expand. ## Similar Debates Have Surrounded Other Technologies This is not the first time digital technologies have triggered heated debates about privacy and data usage. In recent years, controversies flared around facial recognition systems deployed with limited consent. Many experts highlighted how facial recognition algorithms were trained on datasets scraped from social media, online photos, and public cameras without notice or permission. This echoed concerns now arising around ChatGPT consuming personal conversations. However, compared to some past incidents, the response from OpenAI shows greater sensitivity to user perceptions and privacy. Quickly adding options like Incognito Mode illustrates lessons have been learned. ## User Trust and Perception Remains Vital ChatGPT‘s launch has captivated millions worldwide and lent renewed momentum to the AI field. But sustaining engagement long-term requires responsibly nurturing user trust and comfort. If people fear their intimate conversations are feeding systems without their approval, many may disengage. That‘s why moves like Incognito Mode and granular data controls in the Business version are so prudent. ## The Future of Privacy in an AI World As AI capabilities continue rapidly evolving, models will become increasingly adept at analyzing personal user data to optimize performance. This makes transparency and consent a priority now more than ever. Hopefully, the example set by OpenAI in response to privacy pressures precipitated by ChatGPT‘s meteoric rise foreshadows a thoughtful path forward. With great AI power comes great responsibility. Keeping users in the loop must remain central going forward. The ideal scenario provides the benefits of personalization and optimization enabled by voluntarily shared user data, without ever crossing lines that make people uncomfortable. Building a future founded on trust and openness will unlock the full potential of AI while upholding societal values. ## Key Takeaways on the ChatGPT Incognito Mode Rollout - OpenAI introduced Incognito Mode to give ChatGPT users more control over privacy after concerns were raised. - With Incognito Mode, conversations are not saved, linked to accounts, or used for training models. - By default, all ChatGPT conversations are retained to improve the AI via continued learning. - Investigations of ChatGPT in Europe appear to have spurred quicker action from OpenAI. - Upcoming paid ChatGPT subscription expands controls further for professionals. - Balancing personalized AI advancement and privacy protection remains an ongoing challenge. - Increased transparency and consent around data usage helps build user trust. - Thoughtful openness remains crucial as AI systems gain more advanced capabilities. --- Source: [OpenAI Rolls Out Incognito Mode on ChatGPT: Giving Users More Control Over Privacy](https://33rdsquare.com/openai-rolls-out-incognito-mode-on-chatgpt-giving-users-more-control-over-privacy/)