Your ChatGPT history serves as a personality assessment you weren't aware you were undergoing.

      Scientists have developed an AI capable of predicting personality traits from your ChatGPT interactions, and its accuracy is quite impressive.

      Whenever you ask ChatGPT to draft an email, discuss a personal issue, or search for symptoms, you may be sharing more than just a simple question. According to a report by TechXplore, researchers at ETH Zurich created an AI model that successfully predicts personality traits using actual ChatGPT conversation logs.

      The study comprised 62,090 genuine conversations from 668 ChatGPT users, who also completed a standard personality test, providing a reference for measurement. The AI was trained to categorize each participant as low, medium, or high in five different personality traits: openness, conscientiousness, extraversion, agreeableness, and neuroticism.

      The refined model outperformed random chance in identifying all five traits, with extraversion being the most accurately predicted trait, showing a 44% improvement over mere guessing.

      Does the content of your conversations matter?

      The findings revealed that chats focused on mental health made it particularly easy to infer extraversion. Conversations about religion were significantly linked to conscientiousness predictions, while dialogues concerning mental states and moods increased the predictability of openness.

      According to researchers Derya Cögendez, Verena Zimmermann, and Noé Zufferey, even casual discussions contained sufficient information to be valuable. They also observed that the more frequently you use ChatGPT, the easier it is to create a profile of you. Given the extent of data shared with ChatGPT, the ability to identify our personality traits becomes quite significant.

      Recently, ChatGPT has begun incorporating advertisements. With its knowledge of user data, it's concerning to consider how effectively it could tailor ads to influence our thoughts.

      Why is this important beyond academic research?

      The researchers emphasize the implications of their findings. Service providers already possess much of this data, and with over 800 million ChatGPT users monthly as of January 2026, the potential scale for profiling is tremendous.

      A personality profile derived from your chat history could be employed for targeted advertising, personalized manipulation, or, in worst-case scenarios, widespread influence campaigns.

      For now, it is essential to remember that your AI chatbot is not a private journal. You can take control by periodically deleting your ChatGPT history to erase your personal conversations from its memory.

      Rachit is an experienced tech journalist with more than seven years of experience in the consumer technology sector.

      ChatGPT for Excel and Google Sheets has officially exited beta and is a significant development.

      You no longer need to be an Excel expert; ChatGPT will handle the more difficult tasks for you.

      ChatGPT for Excel and Google Sheets is now publicly available after moving out of beta. OpenAI announced the Excel beta in March and the Google Sheets beta in April. With its public release, this is a notable advancement for anyone who frequently utilizes spreadsheets in their work.

      So, what functionality does it offer?

      Microsoft has made significant improvements to Copilot Cowork, making it capable of reducing your to-do list.

      Microsoft aims for its AI to progress beyond merely answering questions to actually initiating and completing tasks on your behalf. This is the essence of Microsoft Copilot Cowork, which was introduced through the Frontier program earlier this year.

      Since its launch, it has acquired various features that enable users to manage their email inboxes, conduct research, generate documents, and create web pages. Recently, CEO Satya Nadella announced a new set of capabilities that significantly enhance Cowork's functionality.

      ChatGPT’s updated default model is now less likely to provide misleading information in medical and financial contexts.

      GPT-5.5 Instant has replaced GPT-5.3 Instant as ChatGPT's default model, with the update rolling out to all users today. The new model aims to enhance three key areas: accuracy, brevity, and personalization.

      Less hallucination, more substance.