> ## Content Index > Fetch the complete content index at: https://www.regenraygun.com/llms.txt > Use this file to discover other available public pages before exploring further. # Tencent shipped AI in League of Legends while Gamescom awarded itself points for abstinence - URL: https://www.regenraygun.com/tencent-shipped-ai-in-league-of-legends-while-gamescom-awarded-itself-points-for-abstinence/ - Published: 2026-09-08T08:59:10.000Z - Updated: 2026-09-08T08:59:10.000Z - Description: The world's largest games company used Gamescom to launch an AI middleware business. - Author: Jon Jordan - Tags: Tencent The most consequential AI announcement at Gamescom 2026 was not made on the Opening Night Live stage, was not covered by anyone who mattered, and contained no controversy whatsoever. On 27th August, [Tencent Games Central Tech](https://centraltech.tencent.com/?ref=regenraygun.com) made its Gamescom debut as a unified brand and laid out six proprietary AI technologies spanning the entire game production lifecycle. Nobody screamed about it. There was no backlash, no review-bombing, no Kotaku list. Meanwhile, roughly a hundred metres away, the Western games industry was conducting an elaborate public morality contest about whether using a machine to generate a placeholder texture makes you a bad person. The [six technologies were presented](https://www.prnewswire.co.uk/news-releases/tencent-games-central-tech-makes-its-gamescom-debut-accelerating-game-industry-innovation-with-full-pipeline-proprietary-ai-technologies-302861485.html?ref=regenraygun.com) as a portfolio covering content production, gameplay and long-term live operations. **Motus AI** is the centrepiece. Built on what Tencent describes as a proprietary family of generative animation foundation models, it connects the entire 3D character animation pipeline: rigging, intelligent skinning, motion generation, motion refinement and real-time character interaction. It produces animation from text prompts, video or keyframes. It handles multi-character interactions including combat, cooperative actions and social scenes. It drives facial expression and body movement from text or audio. And it automatically detects and corrects the specific failures that eat animator time — jitter, clipping, abnormal joint rotation and foot sliding. **GIGA** treats games and virtual worlds as research environments for general-purpose AI agents, working on visual decision-making, language-action alignment and real-time competitive decision-making. Tencent frames the goal as moving foundation models from understanding context to taking action in dynamic environments. **MagicDawn** combines rendering with AI and cloud computing. The case study shown at Gamescom was *Roco Kingdom*, rebuilt from a classic browser game into a 64-square-kilometre cross-platform open world with consistent visual quality across devices. **Anti-Cheat Expert** covers anti-cheat, client hardening, in-game economic risk management and threat intelligence. **GVoice** handles low-latency voice with AI noise reduction, a proprietary AI codec and real-time speech translation. **WeTest Acorn AI Studio** puts AI agents into builds to play them like players and file bugs like QA specialists, across multiple genres and platforms. Tencent also ran four technical sessions at gamescom dev, with titles aimed squarely at technical directors rather than press: on unified AI pipelines for 3D character animation, autonomous AI agents for FPS battlefields, high-performance global illumination, and scalable autonomous testing for complex games. But this isn't aspiration. Motus technology is already deployed in *Game for Peace*, *PUBG Mobile*, *Delta Force*, *Honor of Kings*, *Roco Kingdom*, *League of Legends* and *Teamfight Tactics*. Read that list again. *Honor of Kings* is among the highest-grossing games ever made. *League of Legends* and *Teamfight Tactics* are Riot titles, and Riot has been wholly owned by Tencent since 2015\. Which means AI animation tooling is already shipping inside one of the most-played competitive games in the Western world, has been for some time, and not one person has demanded a Steam disclosure tag for it. No press release was issued. No trailer was accused of being generated. No forum thread went viral. It simply happened, at industrial scale, in games hundreds of millions of people play, while the discourse was busy elsewhere. On the same day Tencent's release went out, TheGamer published an audit of all 55 games shown at Gamescom Opening Night Live. Thirteen of them — 23.6% — disclosed some AI use. Here is what the Western games industry got caught doing. *Train Sim World 7* used text-to-speech for station announcements. *Once Human* used AI in marketing materials. *1666: Amsterdam*, *Crimson Desert: Enhanced Edition* and *The Blood of the Dawnwalker* used AI for placeholder assets and placeholder voice-over, with developers hurriedly assuring everyone the placeholders would be removed before launch. *Aion 2* generated some card designs and icons. *Crazy Taxi: World Tour* used AI for reference material during development. Station announcements. Placeholder textures. Reference images. A total scandal. The response to being caught doing almost nothing was not to shrug, but to compete for abstinence points. Sega said that *Crazy Taxi: World Tour* would contain no AI-generated content in the final release. Panache Digital confirmed that AI elements spotted in *1666: Amsterdam*'s early access build would appear in neither this version nor the finished game. Amplitude spent a day denying that the *Humankind II* trailer was AI-generated, explaining that it was shot at Sofia's Nu Boyana Film Studios with real actors, custom props and blue-screen compositing, while notably declining to say anything about post-production. Kotaku ran a list of Opening Night Live games readers could safely ignore on the grounds of generative AI slop. PC Gamer reported a promising indie getting mobbed over minor AI use, in a week when more than 300 Steam releases carried AI disclosures. TheGamer called its own 23.6% figure generous, and cited a study putting actual studio AI usage at 50% while only 20% disclose it. So the moral high ground is substantially fictional. Half the industry is using this technology. A fifth of the industry admits it. The gap between those numbers is not abstinence; it is public relations. What Gamescom demonstrated was not a principled Western refusal to use AI. It was a Western consensus that using AI is something to be caught at rather than something to be good at. The result is the worst of both positions. Western studios are using AI for the most trivial and least valuable tasks imaginable, apologizing when found out, and promising to stop. Tencent is using it for load-bearing production infrastructure across the largest games portfolio on the planet, and running training sessions to teach other developers how. The difference on display at Gamescom was not ethical. It was competence.