1666 Developer Says AI Was Only Used During Conception, Now Removed

What changed

Patrice Désilets has drawn a new line under the generative AI controversy surrounding 1666: Amsterdam: according to the Assassin’s Creed co-creator, AI is no longer part of the game’s development pipeline. Speaking to GamesRadar+ at Gamescom 2026, Désilets said the studio had used “some AI” mostly “at the beginning,” specifically during “conception and pre-production,” before adding that it has been “totally removed” and that Panache “won’t be using any AI anymore” [1].

That clarification matters because the debate around 1666 was not simply about whether a finished asset had been generated by a tool. It was about trust: players had already scrutinized the game’s playable prologue after its reveal and accused some visible assets of bearing the hallmarks of generative AI. PC Gamer reported on August 31 that Panache had previously acknowledged AI use in the production of some assets, apologized, and promised to remove those elements before the early access launch [2].

Désilets’ latest statement is therefore both a production update and a reputational repair effort. The practical claim is direct: the current development of 1666 is no longer using AI. The more delicate claim is philosophical: that early-stage use in conception and pre-production can be separated from the final authored work. That is where the story remains contentious, even after the studio’s removal pledge.

The timeline behind the backlash

The renewed attention comes after 1666: Amsterdam resurfaced in a more public way in 2026. The game was presented at Summer Game Fest, where its witch-hunt setting, historical premise and Désilets’ long association with Assassin’s Creed gave it immediate visibility [2]. But the positive response shifted when players examined the playable prologue and found assets they believed were AI-generated [2].

GamesRadar+ described the situation as one in which the accusations “turned out to be true,” while also reporting that Panache had committed to removing AI from 1666 before making it playable in early access [1]. Désilets then used Gamescom 2026 to clarify what he says is the game’s present relationship with generative AI: early use, no ongoing use, and no further AI planned [1].

Insider Gaming summarized the same update on August 29, writing that Désilets clarified AI had been used in conception and pre-production, but that it had been completely removed and would not be used going forward [3]. That makes the studio’s current position fairly unambiguous: whatever AI-assisted material existed earlier is not supposed to remain part of the active build or production process.

AFP, in an interview published by Kuwait Times on August 29, added one important detail: Désilets said AI used in pre-production had “ended up in the prologue,” but that all those elements had since been removed and the team no longer uses it [4]. That explains why the controversy became visible to players in the first place. If pre-production material migrated into a public-facing prologue, then the studio’s internal boundary between temporary concept work and shippable content had already failed once.

Why “only in pre-production” is not a simple defense

Désilets’ wording is careful. He is not saying AI was never used. He is saying it was used early, then removed. For some players, that may be enough: if AI-generated elements are gone from the game and the studio has stopped using the technology, the matter is functionally resolved. For others, the phrase “conception and pre-production” is precisely what makes the admission sensitive.

Pre-production is where a game’s identity is formed. It is where visual direction, mood, narrative tone, environmental references and character ideas begin to cohere. PC Gamer’s analysis argued that saying AI was used “only” at the concepting stage does not necessarily make the issue smaller, because this is often the most imaginative and formative phase of development [2].

That concern is not merely aesthetic. When a studio uses generative AI to explore concepts, the tool may influence what the team sees as possible, appealing or efficient. Even if the final textures, props, images or written elements are later replaced by human-made work, early AI outputs may still have affected the direction of the project. The question is not just whether an AI-generated file remains in the build; it is whether the project’s creative DNA was shaped by machine-generated references.

Désilets’ current answer is to emphasize removal and future non-use. That answer addresses the practical layer of the controversy. It does not fully settle the wider industry debate over whether concept-stage AI is ethically different from final-asset AI. In 1666’s case, the debate is sharper because the game is being sold partly on authorship: a personal, long-delayed project from a designer whose reputation is tied to authored historical worlds.

The early access context

1666: Amsterdam entered early access on PC on August 25, 2026, according to Insider Gaming, which also reported an introductory price of $24.29 [3]. AFP reported that the first two chapters had been released for PC in early access and that the final version is expected to include around a dozen chapters, according to Désilets [4].

That early access framing matters. The game is not being presented as complete. Panache is inviting players into a version of the project that is still under development, with more chapters and refinements to come. In that context, the AI removal pledge becomes part of the development roadmap even if it is not framed as a feature. Players are being asked to judge the current build while also trusting the studio’s commitment about how future content will be made.

PC Gamer reported on August 31 that, at that moment, the Steam page did not show an AI disclosure and that user reception appeared divided, with a “Mixed” rating and 66% positive reviews at the time of its report [2]. Those figures may move as more players buy and review the game, but the snapshot captures the launch-week environment: curiosity about the setting and design, mixed reactions to execution, and unresolved skepticism around the AI controversy.

Désilets told AFP he was “very satisfied” with initial feedback, while no sales figures or expectations had been released [4]. That gives Panache some room to rebuild trust, but not unlimited room. Early access players tend to tolerate rough edges when they believe a studio is transparent. The AI issue has made transparency a central requirement for 1666 rather than a secondary communications concern.

A personal project under unusual scrutiny

The AI controversy also lands on a game with an unusually long and personal development history. AFP reported that Désilets said he first had the idea for 1666: Amsterdam during a trip to Amsterdam in spring 2011, after leaving Ubisoft [4]. The project later moved through THQ Montreal before that studio’s bankruptcy and Ubisoft’s acquisition, followed by legal disputes that ended with Désilets regaining the rights in 2016 [4].

That history explains why the game’s return carried emotional weight. Désilets told AFP that the Summer Game Fest moment marked the end of a “very personal” battle, and the report described 1666 as a project shaped by more than 15 years, a bankruptcy, and legal conflict [4]. The AI backlash cuts against that narrative because the appeal of 1666 is tied to the idea of a hard-won creative vision finally returning to life.

The game itself is a dark, story-led action adventure set around European witch hunts, with players controlling Noa, a witch or demon hunter figure, across different eras including 1666, 1999 and the present [4]. GamesRadar+ also highlighted its unusual premise, including Noa’s sorcery and a playable cat companion connected to 1999, framing the game as an eccentric, historically inflected project rather than a generic action title [1].

That specificity is why the AI question has been so damaging. A game about witchcraft, historical paranoia and hidden identities depends heavily on atmosphere. If players suspect its images or environments are built from generic AI outputs, the spell breaks. Panache’s challenge is not only to remove files; it must persuade players that the game’s atmosphere now comes from a human creative team.

What Panache still needs to prove

The most important current fact is clear: Désilets says AI was used only at the beginning, in conception and pre-production, and is now removed from 1666’s development [1]. Multiple reports published over the past three days repeat the same core position: Panache says it is no longer using AI and has removed the elements that triggered the controversy [2][3][4].

What remains less clear is how the studio will document that change. Will it publish a more detailed production note? Will it identify the categories of assets that were replaced? Will future updates include an explicit statement about AI use or non-use? The available reports establish Désilets’ commitment, but they do not describe an independent audit, a full asset list, or a formal public policy beyond the promise that AI is gone.

For players who simply wanted the offending assets removed, Désilets’ statement may be enough. For players worried about labor, attribution and the role of generative tools in creative work, 1666 remains a case study in how quickly “temporary” AI can become public-facing content. The prologue incident shows that pre-production experiments are not always safely contained.

The best path forward for Panache is consistency. If future early access updates arrive without renewed AI questions, the controversy may recede into the game’s difficult launch history. If players find anything that appears to contradict the studio’s promise, the backlash will likely return with greater force. Désilets has now made the commitment in plain language: conception and pre-production only, removed now, no AI going forward [1].

For 1666: Amsterdam, the story is no longer whether AI was used. It was. The story now is whether Panache can convince players that the game they are being asked to support in early access is, from this point forward, made without it.

Sources from the last 72 hours

1. [1]After AI backlash, Assassin's Creed creator says 1666: Amsterdam "won't be using any AI anymore" and only used it "in conception and pre-production"Aug 28, 2026, 8:21 PM UTC
2. [2]1666: Amsterdam's developer says it 'won't be using AI anymore', and the tech was only used 'in conception and pre-production'—arguably the most creative phase of the entire processAug 31, 2026, 12:00 AM UTC
3. [3]Assassin’s Creed Creator Promises to No longer Use AI in 1666: Amsterdam’s DevelopmentAug 29, 2026, 12:00 AM UTC
4. [4]For ‘Assassin’s Creed’ creator, new game ends a ‘very personal’ battleAug 29, 2026, 5:26 PM UTC

AI-generated article based on recent web research, then preserved as a dated editorial snapshot.