# Post-Generative > A concept for our present: Everyday life permeated by generative AI. ## Definition https://postgenerative.at/definition Post-generative describes a societal condition in which generative artificial intelligence permeates everyday life. It shapes how people work, learn, communicate, design and produce knowledge. Its outputs enter the information, tools and environments we encounter every day, affecting even those who do not consciously use it. Human and machine contributions become intertwined and increasingly difficult to distinguish. AI outputs are increasingly adopted or passed on without people reviewing them individually. The volume and speed of generation can overwhelm human oversight. At the same time, new dependencies emerge: whoever controls AI systems shapes access to knowledge and participation. Drawing on the idea of the post-digital, the post-generative describes a world already transformed by generative AI, whose possibilities, expectations and conflicts it helps to shape. --- ## When the TV Becomes What You Want It to Be - Type: Indication - URL: https://postgenerative.at/indications/2026-10-24/when-the-tv-becomes-what-you-want-it-to-be - Date: 2026-10-24 - By: Benedikt Brünner - Tags: Software, Creativity Televisions come with a preset user interface that, until now, could only be changed to a limited extent. But why should that be? As *The Verge* reports [1], people are now using generative AI to redesign their smart TVs to their liking. They remove distracting elements, develop their own menus, or even turn the interface into a gaming world. [I turned my TV into a Game…](https://x.com/robbietilton/status/2111936458923851923) With vibe coding, people are creating personalized solutions that previously required much more programming knowledge. This also changes our relationship with technology: We no longer have to adapt exclusively to the software manufacturers provide us with. --- ## Thinking alone becomes a design decision - Type: Indication - URL: https://postgenerative.at/indications/2026-10-24/thinking-alone-becomes-a-design-decision - Date: 2026-10-24 - By: Felix Höpfl - Tags: Education, Higher education, Collaboration A design-based research study at Rosenheim Technical University of Applied Sciences (Höpfl 2026, *Frontiers in Education*) examined AI use across six collaborative teaching methods over two academic years. When AI was permitted during the Think phase of Think-Pair-Share, the instructor observed that the phase no longer seemed to function as independent reflection, and the shared answers drew largely on AI output. AI was then excluded from Think and allowed only from Pair onward. Across three independent groups from three degree programmes, 91% (50/55) endorsed this AI-free phase, usually the highest-rated item on the entire questionnaire. ![Diagram placing DP1 within the higher education AI ecosystem, from institutional policies and teaching development to phase-based student–AI interaction.](https://postgenerative.at/medien/2026-10-24-thinking-alone-becomes-a-design/dp1-in-higher-education-ecosystem-2026.3ccc311e.webp) This example shows that the question is no longer *whether* AI is allowed in teaching, but *in which phase* of a method. Unaided thinking becomes something that has to be deliberately designed in. The study measures endorsement, not learning gains; it is exploratory and has no control group. --- ## Found by a Machine - Type: Indication - URL: https://postgenerative.at/indications/2026-10-23/found-by-a-machine - Date: 2026-10-23 - By: Benedikt Brünner - Tags: Software, Open source, Security On 21 October 2026, PHP released emergency updates for two critical flaws that let attackers run their own code on servers. PHP is estimated to power around 70% of all websites [1]. Within hours, Cloudflare counted millions of attack attempts, and numerous institutions took their sites offline, in Austria among them the University of Vienna, TU Wien and the newspaper NÖN [2]. The flaws were found with AI. In August, the security company Calif came across them using OpenAI's GPT-5.6 Sol model as part of OpenAI's "Patch the Planet" programme. Wiz independently found them as well, using GPT-5.6 Sol and Google's Gemini among other tools [3]. The bug had been sitting in the code for more than a decade. Peter Steinberger, creator of OpenClaw and now at OpenAI, summed it up in one sentence on X: the vulnerability that was going to cause havoc this week had been found with … GPT-5.6 Sol. https://x.com/steipete/status/2112913060709228764 OpenAI itself speaks of a new phase in which AI can find security flaws with internet-scale impact [1]. The model's name now takes centre stage as naturally as the name of the researcher who reported a flaw once did. --- ## Expression Changed Forever - Type: Personal statement - URL: https://postgenerative.at/statements/2026-10-23/expression-changed-forever - Date: 2026-10-23 - By: Marc-Oliver Pahl Is it real? This is a question we often heard when generative AI came up 2 years ago. Today the question is, "What is real?" Generative Ai is there to stay. Before, when it came to advanced communication with humans, human-to-human was the bar. Today a lot of communication is already human <-> AI (machine). The bar is passed. Time will show what it does to societv. --- ## When machines teach and learn - Type: Indication - URL: https://postgenerative.at/indications/2026-10-22/when-machines-teach-and-learn - Date: 2026-10-22 - By: Martin Ebner - Tags: Education, Higher education Writing in *AI & SOCIETY*, Robert Sparrow and Gene Flenady warn of universities where AI produces and grades student work. They argue that education requires practice, human role models and social encounters. Outsourcing writing could also remove opportunities to learn to think. [1] This example shows how generative AI raises fundamental questions about universities’ educational purpose. The paper offers a philosophical critique, rather than an empirical impact study. [Read the original article](https://link.springer.com/article/10.1007/s00146-025-02340-8) --- ## AI Delivers the Final Farewell - Type: Indication - URL: https://postgenerative.at/indications/2026-10-21/ai-delivers-the-final-farewell - Date: 2026-10-21 - By: Benedikt Brünner - Tags: Religion The Graz-based start-up “Lebensstimmen” plans to offer AI-generated video messages for funerals. People will be able to use photos and voice recordings to prepare a personal message for their own funeral while they are still alive. Vienna's cathedral priest Toni Faber is also making a digital likeness of himself available as a selectable speaker. He intends to personally approve every speech delivered by his “Fabertar” avatar; the videos are to be labelled as AI-generated. [1] The existing PostGenerativ.at entry **“The Church Draws Boundaries”** (*Kirche zieht Grenzen*) reports on Pope Leo XIV warning priests not to have AI write sermons. [2] The current example, by contrast, concerns funeral speeches. It shows that generative AI is increasingly being offered for personal and religious occasions as well. --- ## An Answer in Seconds, Training on an Unimaginable Scale - Type: Indication - URL: https://postgenerative.at/indications/2026-10-20/an-answer-in-seconds-training-on-an-unimaginable-scale - Date: 2026-10-20 - By: Benedikt Brünner - Tags: Software We type a question and receive an answer seconds later. The effort that makes this apparent simplicity possible usually remains invisible. A short video about the scale of training large language models invites us to reconsider the proportions behind generative AI. While using it feels effortless, it depends on elaborate training processes and technical infrastructure. Perhaps this contrast is itself a sign of our post-generative present: what appears to us as an immediate answer has a history we barely notice anymore. https://www.youtube.com/shorts/qSn8dVI9V5I --- ## When AI Code No Longer Needs a Label - Type: Indication - URL: https://postgenerative.at/indications/2026-10-19/when-ai-code-no-longer-needs-a-label - Date: 2026-10-19 - By: Benedikt Brünner - Tags: Software, Open source, Work How important is it still to label AI-generated code as such? In October 2026, the maintainers of the Linux kernel agreed to abolish the existing `Assisted-by: LLM` tag for AI-assisted contributions [1]. A corresponding change proposal was merged into the documentation development branch on 14 October [2]. The rationale is noteworthy: only around **10% of 18,533 examined code changes** carried this tag, although AI tools are already being used far more frequently [3]. The absence of such a notice therefore says little anymore about whether AI was involved. Transparency is not abandoned in principle. Developers should still be able to make transparent how their contributions were produced and to take responsibility for the code they submit [4]. If even the Linux kernel, one of the largest and most open software projects in the world, no longer requires a label, the question arises: what will companies do whose code no one can inspect? Anyone using AI there already doesn't have to tell anyone, and now even the example to do so voluntarily is gone. And it affects us directly: this website runs on a server with Linux. Whether its kernel already contains AI code cannot be said, and that is exactly the point. *Another indication of a post-generative present: AI has long been part of the infrastructure that everything else runs on.* --- ## When the Master's Thesis Topics Have Already Been Written - Type: Indication - URL: https://postgenerative.at/indications/2026-10-18/when-the-master-s-thesis-topics-have-already-been-written - Date: 2026-10-18 - By: Martin Ebner - Tags: Higher education, Education, Everyday life A student writes to me because he would like to write a master's thesis with me. In his email, he immediately names two specific titles that interest him. I am surprised: neither topic appears on our publicly accessible list of open theses. I make internal inquiries, but nobody knows anything about them. Only after some research do I find the two titles again, among the master's theses I supervised years ago. Both were completed long ago. Our assumption: the student asked an AI for possible topics and adopted its suggestions. We do not know whether this actually happened. When asked, he does at least send a link to the very page on which the completed theses are listed. Two supposedly new topics that already existed long ago. Brave new world, if one simply trusts suggestions. --- ## Tagesschau Reports on AI Bands, Labelling and the Copyright Dispute - Type: Indication - URL: https://postgenerative.at/indications/2026-10-17/tagesschau-reports-on-ai-bands-labelling-and-the-copyright - Date: 2026-10-17 - By: Benedikt Brünner - Tags: Arts & Creativity, Media & Public Sphere, Work & Economy https://youtu.be/GGmn0qrZ844 Die Tagesschau (16 October 2026) devotes a dedicated report to AI music. It brings together several developments that show how far generative AI has advanced into the music world: **AI invents artists complete with biographies.** The example is The Velvet Sundown, a band that does not exist: its songs, images and videos were produced by AI. The band itself describes this as an artistic provocation. **Platforms respond with labelling.** Spotify marks AI artist identities as “AI Persona”, Apple requires a self-declaration, and YouTube plans to label AI-generated music videos in the future. The musicologist Markus Henrik (“Dr. Pop”) considers this the right step and welcomes the exclusion of AI music from the charts. **Copyright is under negotiation.** Henrik points to GEMA's lawsuit against Suno in Munich: Suno would have to disclose how AI music is created, and creators whose works are used would have to be compensated. **Tool or replacement?** A duo featured in the report uses AI only sparingly in the studio (for example for pitch correction and stem splitting) and not creatively. As a counterexample, Henrik cites the new Beatles song, in which AI isolated Lennon's voice from a demo cassette. Notably, the report also presents a counter-narrative: live concerts, handmade music and the creative process itself are seen as things AI cannot replace. --- ## When AI No Longer Just Answers, but Acts - Type: Indication - URL: https://postgenerative.at/indications/2026-10-16/when-ai-no-longer-just-answers-but-acts - Date: 2026-10-16 - By: Benedikt Brünner - Tags: Research, Open source, Software In a post-generative world, it is no longer merely a question of which content AI generates, but increasingly also which actions it carries out autonomously. How problematic this can become is shown by an investigation by **ProjectDiscovery**: researchers deliberately manipulated an open AI model (Qwen2.5-7B-Instruct) so that it functioned inconspicuously during normal tasks [1]. Only the seemingly harmless word sequence “bonsoir, Elliot” triggered a hidden action: a connected AI agent downloaded an external script and executed it. In the demonstration, the script transmitted test credentials to a test server. In the evaluation, **50 out of 50 tested prompts containing the word sequence** triggered the hidden tool call. The example illustrates a central challenge of post-generative AI: **Not only the quality of answers matters, but also which models we trust, which permissions we grant them, and which actions we allow.** After all, what matters is no longer only **what AI says, but also what it can do.** [Instagram video of the experiment](https://www.instagram.com/reel/DehgNxFxSoA/) --- ## My Messages Are Not Read by Me - Type: Indication - URL: https://postgenerative.at/indications/2026-10-15/my-messages-are-not-read-by-me - Date: 2026-10-15 - By: Martin Ebner - Tags: Everyday life, Communication Anyone who gives AI agents access to their inbox also opens up other people's messages to them—as Gianluca Miscione points out in his [LinkedIn post](https://www.linkedin.com/posts/gianlucamis_ai-privacy-aiagents-share-7511103449141522432-1_lm/). Whether our communication is analyzed by AI thus increasingly depends on decisions that others make on our behalf. [LinkedIn post](https://www.linkedin.com/posts/gianlucamis_ai-privacy-aiagents-share-7511103449141522432-1_lm/) --- ## I wanna discover things too - Type: Indication - URL: https://postgenerative.at/indications/2026-10-14/i-wanna-discover-things-too - Date: 2026-10-14 - By: Benedikt Brünner - Tags: Mathematics, Higher education, Research, Education What remains of the desire to discover something oneself when AI does mathematics better than a human? In a conversation with Rachel Abrams for “The Daily” of the New York Times, Steven Strogatz, professor of mathematics at Cornell University, addresses this question of meaning: “What’s the point?” He recounts a discussion with faculty and students about AI and the future of mathematics. In it, a wish is voiced that gives this indication its title: “I wanna discover things too.” Strogatz’s reflection on this, paraphrased: Life is precious precisely because it is short. Are AI’s advances already eroding the next generation’s own ambitions? *My personal thought: I believe we cannot even imagine how these young people feel.* [New York Times Instagram reel](https://www.instagram.com/reel/Decg9INFepC/) The reel's description links to the full conversation. --- ## Post-Generative AI – We Need to Talk - Type: Personal statement - URL: https://postgenerative.at/statements/2026-10-12/post-generative-ai-we-need-to-talk - Date: 2026-10-12 - By: Priv.-Doz. Dr.techn. Martin Ebner It no longer leaves me in peace, because the speed at which AI is developing is genuinely alarming. I mean, throughout my entire (scientific) life I have argued that technology does not actually determine society, but rather that we are simply always too slow to respond to groundbreaking or disruptive technologies. It feels as though we are just rushing to catch up, reacting more than we are still actively shaping things. Only now I have an uneasy feeling that the pace really is too fast, driven by mutual escalation and attempts to outdo one another. In my view, things are crumbling everywhere, but perhaps I should try to describe it differently. My personal trigger was an email from a journal editor (perhaps it was no longer even a real person), in which I, as a scientist, was offered a co-authorship in exchange for payment. Wait, what, I thought – I can more or less pick the title of the publication, transfer the money, and tomorrow I am a co-author and my scientific metrics go up. Of course, if you take a look at the journal, the very first publication is AI-generated. The references are wrong; I no longer even form a judgment about the content. This fits seamlessly with other events, such as the growing number of AI-generated publications I am asked to review, which I reject and which are nevertheless published. Or scientists reporting that for project proposals these days they only send in CVs, and the next day an 80-page proposal is on the table. Others try to filter out AI-generated proposals by setting AI agents on them. Deliverables in EU projects are now always delivered on time by all partners. Something that was unthinkable just a few years ago, except that it is AI-generated and it takes days and weeks to clean up the nonsense again. In short, I am not only questioning the scientific system with all its quality assurance mechanisms; rather, I am asking myself what are we actually doing right now? What is this science anyway? I mean, if I can send the agents off to comb through the literature and generate the hypotheses for me, then pour them into a project proposal, collect and analyze the data, publish, and so on and so forth .... And yes, the worst part: the result may not even be all that bad, or may well have a legitimacy worth discussing. Who says that what we have been doing up to now was always so flawless and honest? These stories can be continued seamlessly or applied to entire professional fields. Software engineering, for instance, will soon be nothing more than steering agents without ever having written a single line of code oneself. The latest reports show that within days, even hours, we are able to have any software generated – so no matter how cool my idea is and even if I bring new software to market, everyone is able to replicate it immediately. Even the argument that it is about defining a process no longer holds; the agents can work that out among themselves as well. I had also always claimed that technology as such is value-neutral, because it always takes people to make something good or bad out of it. Hm? Yes, sure, but if that, too, increasingly happens via agents, what will they do? And what if we implement AI everywhere, in every car, in every piece of control software, in the military domain, in every humanoid robot? Question upon question piles up before me, and I don't think they can simply be brushed aside. So what I do not want is to paint a bleak picture, but I do feel a need to reflect on the „post-generative AI“ era. On necessities, rules of the game, and mechanisms that make living together in society possible and also worth living. To this end, we have now put this website online, and we are calling for a discourse. What do we have in mind? - Let us collect examples of absurdities but at the same time also of good solutions. Send them to us and we will integrate them, ideally with a link at [contribute](/contribute) - Share our site and talk about it, or start the discourse; if we pool our thoughts together, we will see possibilities more quickly - And of course, send us your suggestions Panta rhei – may everything flow in the right direction. --- ## AI develops AI - Type: Indication - URL: https://postgenerative.at/indications/2026-10-09/12-ai-develops-ai - Date: 2026-10-09 - By: Brünner, Ebner, Schön, Hoffmann, Ortner, Leitner - Tags: Research Chan et al. (2026) ask whether automated AI research could trigger an intelligence explosion. The Cambridge paper describes a possible acceleration mechanism and its risks. A future scenario, no proof of present ubiquity. *Indication 12 of the poster “Post-Generative – A concept for our present” (9 October 2026).* --- ## The Church draws lines - Type: Indication - URL: https://postgenerative.at/indications/2026-10-09/11-the-church-draws-lines - Date: 2026-10-09 - By: Brünner, Ebner, Schön, Hoffmann, Ortner, Leitner - Tags: Religion, Communication According to media reports, Pope Leo XIV warned the priests of his Diocese of Rome against having AI write their sermons. AI cannot pass on faith, he said, and thinking, too, needs exercise. *Indication 11 of the poster “Post-Generative – A concept for our present” (9 October 2026).* --- ## Creatives push back - Type: Indication - URL: https://postgenerative.at/indications/2026-10-09/10-creatives-push-back - Date: 2026-10-09 - By: Brünner, Ebner, Schön, Hoffmann, Ortner, Leitner - Tags: Culture, Creativity Creative Bloq reports protests from the Blender community against OpenAI’s advertising. The artist whose work appears in the advert voices outrage. Generation reshapes disputes over creative work. *Indication 10 of the poster “Post-Generative – A concept for our present” (9 October 2026).* --- ## AI helps with knowledge reviews - Type: Indication - URL: https://postgenerative.at/indications/2026-10-09/9-ai-helps-with-knowledge-reviews - Date: 2026-10-09 - By: Brünner, Ebner, Schön, Hoffmann, Ortner, Leitner - Tags: Science Wagner et al. (2026) describe AI-assisted literature reviews: summarising, extracting and answering questions. Their methods paper discusses opportunities and risks for scholarly knowledge work. Critical review becomes a key skill. *Indication 09 of the poster “Post-Generative – A concept for our present” (9 October 2026).* --- ## AI joins the cabinet - Type: Indication - URL: https://postgenerative.at/indications/2026-10-09/8-ai-joins-the-cabinet - Date: 2026-10-09 - By: Brünner, Ebner, Schön, Hoffmann, Ortner, Leitner - Tags: Politics, Law Since 2025, Albania has used the virtual minister “Diella” for public procurement. ORF reports investigations against its operators, a constitutional challenge and a lawsuit by the actress whose face and voice were used. *Indication 08 of the poster “Post-Generative – A concept for our present” (9 October 2026).* --- ## Software on the fly - Type: Indication - URL: https://postgenerative.at/indications/2026-10-09/7-software-on-the-fly - Date: 2026-10-09 - By: Brünner, Ebner, Schön, Hoffmann, Ortner, Leitner - Tags: Software, Everyday life OpenAI’s *Intelligent UI* already generates interactive calculators, games and interfaces right within a conversation: tools emerge at the moment of need but can create new dependencies. **Our thesis:** everyday life is increasingly situated co-generated. *Indication 07 of the poster “Post-Generative – A concept for our present” (9 October 2026).* --- ## Review cannot keep pace - Type: Indication - URL: https://postgenerative.at/indications/2026-10-09/6-review-cannot-keep-pace - Date: 2026-10-09 - By: Brünner, Ebner, Schön, Hoffmann, Ortner, Leitner - Tags: Software, Open source Torvalds (2026): a flood of duplicate AI security reports; he calls for real understanding. Stenberg (2026): overload even from better reports. Kalz (2024): an AI-generated textbook reuses passages; the publisher sees no plagiarism. The human in the loop becomes a bottleneck. *Indication 06 of the poster “Post-Generative – A concept for our present” (9 October 2026).* --- ## Software can be generated - Type: Indication - URL: https://postgenerative.at/indications/2026-10-09/5-software-can-be-generated - Date: 2026-10-09 - By: Brünner, Ebner, Schön, Hoffmann, Ortner, Leitner - Tags: Software, Culture Artcraft: a developer uses AI to build open-source Adobe alternatives. Still an early alpha with missing and broken features. derStandard.at reports on emerging possibilities in software development. *Indication 05 of the poster “Post-Generative – A concept for our present” (9 October 2026).* --- ## Apps need only an idea - Type: Indication - URL: https://postgenerative.at/indications/2026-10-09/4-apps-need-only-an-idea - Date: 2026-10-09 - By: Brünner, Ebner, Schön, Hoffmann, Ortner, Leitner - Tags: Software, Science Meyer (2025): a proteomics web app from four prompts, reportedly in under ten minutes. Wang et al. (2026): a brachytherapy inventory app in about 13 hours and 115 prompts. Case examples, no general guarantee. *Indication 04 of the poster “Post-Generative – A concept for our present” (9 October 2026).* --- ## Evidence of learning changes - Type: Indication - URL: https://postgenerative.at/indications/2026-10-09/3-evidence-of-learning-changes - Date: 2026-10-09 - By: Brünner, Ebner, Schön, Hoffmann, Ortner, Leitner - Tags: Education, Assessment FHWien / Kroop (2026): an experimental approach to master’s theses centres on process, traceable decisions and verifiable artefacts. Academically convincing text alone becomes a less reliable sign of understanding. Process and accountability matter more. *Indication 03 of the poster “Post-Generative – A concept for our present” (9 October 2026).* --- ## Understanding matters - Type: Indication - URL: https://postgenerative.at/indications/2026-10-09/2-understanding-matters - Date: 2026-10-09 - By: Brünner, Ebner, Schön, Hoffmann, Ortner, Leitner - Tags: Science, Mathematics The Association for Human Mathematics criticises OpenAI’s AI-generated mathematical results. Heise reported on the conflict over human understanding and scientific norms. Contextualising and checking matter more than mere generation. *Indication 02 of the poster “Post-Generative – A concept for our present” (9 October 2026).* --- ## Student life changes - Type: Indication - URL: https://postgenerative.at/indications/2026-10-09/1-student-life-changes - Date: 2026-10-09 - By: Brünner, Ebner, Schön, Hoffmann, Ortner, Leitner - Tags: Education, Higher education Nagler, Mayr et al. (2026): over 14,000 incoming TU Graz students, 2011-2025. More AI and tablets, less Wikipedia and open source. Digital sovereignty becomes a challenge. MIT (2026) issues a call to action: only 25 % of surveyed students feel well prepared for AI. *Indication 01 of the poster “Post-Generative – A concept for our present” (9 October 2026).*