REDES.@jdonascimentoEspañol
7 Oct 2026 · COLLECTION

Is your work worth less because you used AI?

An invisible watermark can start an uncomfortable conversation. Assessing your work also means understanding your decisions, your knowledge and the rules of the task.

Read related article
Blue background, dark cards and lime and pink accents. Is your work worth less because you used AI? Imagine submitting work you can defend. Someone asks: “Did you use AI?” Would that change how they judge it? Swipe: What does an invisible watermark reveal? Hypothetical situation · Reflection and news @Jdonascimento.
Carousel · 01
Blue background, dark cards and lime and pink accents. OpenAI announced an invisible watermark for its text It is called textGrain It uses a pattern in word choices to help recognise text from OpenAI models. Who is covered by the rollout? It covers eligible ChatGPT and Codex text in the European Union, over the coming weeks. The announcement gives no general rollout date for the UK, Latin America or the US. OpenAI · Announcement, 5 Oct 2026 · Source in caption @Jdonascimento.
Carousel · 02
Blue background, dark cards and lime and pink accents. Could your boss or teacher detect it? What is it for? OpenAI says it aims to make text origins more transparent and respond to European rules. Who gets access initially? Approved researchers and expert organisations will be able to use the detector to evaluate and improve it. According to the announcement, the detector will not be public at launch. OpenAI · Announcement, 5 Oct 2026 · Source in caption @Jdonascimento.
Carousel · 03
Blue background, dark cards and lime and pink accents. The watermark does not measure your contribution What it does not reveal It does not measure human input or identify the user. It does not establish that the text is accurate. The detector can be wrong It may report a mark that is absent or miss one that is present. Unmarked text could still be AI-generated. The result needs context. OpenAI · Announcement, 5 Oct 2026 · Source in caption @Jdonascimento.
Carousel · 04
Blue background, dark cards and lime and pink accents. What did you contribute to the work? Understand what you submit Asking for an explanation to learn and submitting an answer you do not understand involve different contributions. Respect the purpose of the task An exercise may assess a skill you need to practise without AI. A correct result alone does not fulfil that purpose. Using AI alone does not define your ability. Editorial reflection · The examples are hypothetical @Jdonascimento.
Carousel · 05
Blue background, dark cards and lime and pink accents. How to document your contribution 01 Keep a record Save drafts, sources and important changes. 02 Check and explain Check the facts and make sure you can explain your decisions. 03 Check the rules Confirm what use is allowed and disclose assistance where required. Disclosing AI use does not replace permission to use it. Editorial recommendations · Follow the rules for each task @Jdonascimento.
Carousel · 06
Blue background, dark cards and lime and pink accents. Which tasks would you accept AI help with? Think of a specific example It could be a class assignment, a professional report or a social media post. What conditions would you set? What should the person do independently? When would you require work without AI? Share your view in the comments. An open discussion · @Jdonascimento @Jdonascimento.
Carousel · 07
Blue background, dark cards and lime and pink accents. Does using AI change the value of your work? Imagine you can explain every decision. Someone asks whether you used AI. Should AI use count against you? Read the full carousel for context. Hypothetical situation · Edition of 7 Oct @Jdonascimento.
Story · 08
Blue background, dark cards and lime and pink accents. Invisible watermarks: what are they for? A signal about origins OpenAI announced textGrain for eligible ChatGPT and Codex text in the EU, over the coming weeks. A signal with limits It does not measure human input. Its detector can be wrong and will have restricted access initially. The carousel explains who can access the detector. OpenAI · Announcement, 5 Oct 2026 · Source in caption @Jdonascimento.
Story · 09
Blue background, dark cards and lime and pink accents. Where would you accept AI help? Compare two tasks One allows AI assistance. The other asks you to practise a skill independently. Share your view What should the person demonstrate about their work in each case? Comment on the carousel. More content on our website. More content · redes-jdonascimento.pages.dev/en/ @Jdonascimento.
Story · 10
Blue background, dark cards and lime and pink accents. Make your contribution visible 01 Keep a record Save drafts, important changes and sources. 02 Check Review facts, references and conclusions. 03 Explain Make sure you can defend your decisions. 04 Follow the rules Confirm what help is allowed and when to disclose it. Consider the process, understanding and rules of the task. Editorial recommendations · @Jdonascimento @Jdonascimento.
Summary · 11
Blue background, dark cards and lime and pink accents. Is your work worth less because you used AI? Invisible watermarks: what they reveal about your contribution. @Jdonascimento.
Facebook · 12
Blue background, dark cards and lime and pink accents. Is your work worth less because you used AI? Invisible watermarks: what they reveal about your contribution. @Jdonascimento.
X · 13

Instagram

Would you think differently about a good piece of work if you knew its author had used AI? 👀 OpenAI announced this rollout for eligible ChatGPT and Codex text in the European Union. Swipe to discover what new invisible watermarks are for, who will be able to detect them and what they can tell you about your contribution. Which tasks would you accept AI help with? Note: For applications connected to OpenAI through its API, whoever manages the integration can enable watermarking for supported models worldwide. It is initially disabled. Source: OpenAI, 5 October 2026. https://openai.com/index/eu-text-provenance/ For more content, visit: https://redes-jdonascimento.pages.dev/en/

Facebook

Is your work worth less because you used AI? Imagine submitting a report you can explain and defend. Someone asks whether you used ChatGPT. What would they need to know to judge it fairly? OpenAI has announced textGrain, an invisible pattern in word choices. It plans to add it to eligible ChatGPT and Codex text in the European Union over the coming weeks. Initially, detector access will be restricted to approved researchers and expert organisations. The watermark does not measure human input or establish that the text is accurate. The detector can be wrong. To assess the work, we suggest considering what its author understood and checked, the decisions they made and what the task allowed. Keep your drafts and sources; check what help you can use and when you need to disclose it. Which tasks would you accept AI help with, and which would you require people to complete without it? Note: For applications connected to OpenAI through its API, whoever manages the integration can enable watermarking for supported models worldwide. It is initially disabled. Source: OpenAI, announcement dated 5 October 2026. https://openai.com/index/eu-text-provenance/ For more content, visit: https://redes-jdonascimento.pages.dev/en/

X

1/7 Is your work worth less because you used AI? Imagine you can explain and defend what you submitted. Someone asks: “Did you use ChatGPT?” What would they need to know to judge it? Read the thread. 2/7 OpenAI announced watermarks for eligible ChatGPT and Codex text in the European Union, over the coming weeks. The announcement gives no general rollout date for the UK, Latin America or the US. 3/7 textGrain aims to identify text origins. OpenAI says detector access will initially be limited to approved researchers and expert organisations. Announcement, 5 Oct: https://openai.com/index/eu-text-provenance/ 4/7 The watermark does not measure human input or identify the user. It does not establish accuracy. Detection can be wrong; finding no watermark does not prove human authorship. 5/7 Using AI alone does not define your ability. Understanding your work and respecting the task matter: an exercise may require you to practise a skill independently. 6/7 Document your contribution: keep drafts and sources, check the facts and explain your decisions. Check what help is allowed and when to disclose it. Disclosure does not replace permission to use AI. 7/7 Which tasks would you accept AI help with, and which would you require people to complete without it? Share an example. For more content, visit: https://redes-jdonascimento.pages.dev/en/

Share this publication

You may share links, download the resources made available and republish my own material unchanged and for non-commercial purposes. Retain my credit and include a link to the original REDES publication. Third-party resources remain subject to their own terms. Terms of use.