Ubisoft tests gen AI in Far Cry 7, results disappoint
Key insights
- Ubisoft is using Far Cry 7's live development environment as an R&D sandbox for its own in-house generative AI tools.
- Insider Tom Henderson described early AI test results as poor quality, suggesting the tooling remains far from production-ready.
- No generative AI features are planned for the final Far Cry 7 release; the experiments are purely evaluative for future projects.
Why this matters
Major AAA studios using active flagship productions as AI test beds signals that game development pipelines are being restructured around AI evaluation cycles before the tools are mature, creating quality and timeline risk on high-budget titles. Ubisoft's in-house tooling strategy, rather than licensing third-party solutions, means the company is betting on proprietary AI infrastructure at scale while publicly distancing the experiments from the shipped product. The insider's negative quality assessment and the deleted post together illustrate the reputational pressure studios face when AI adoption becomes visible to the public, which will shape how the entire games industry communicates about AI integration going forward.
Summary
Ubisoft is quietly running generative AI experiments inside Far Cry 7's development environment, and according to gaming insider Tom Henderson, the early results are not promising. The tests use text prompts fed into the Far Cry 7 game world to stress-test Ubisoft's in-house AI tooling, with the stated goal of evaluating what those tools might be capable of on future projects.
The experiments are framed explicitly as R&D sandboxing, not features destined for the shipped game. Henderson deleted his original post after it circulated, then confirmed the work is internal-only and no generative AI features are planned for Far Cry 7's release.
Essentially: Ubisoft is using an active AAA production as a live test bed for AI tooling that isn't ready.
- The evaluation method feeds natural-language prompts directly into the Far Cry 7 environment to measure output quality from Ubisoft's proprietary tools.
- Henderson's characterization of the results as looking bad suggests the tooling is pre-production grade, not close to shipping quality.
- The deletion of the original post points to sensitivity inside Ubisoft about public perception of AI use in major franchises.
Game studios are increasingly caught between investor pressure to adopt AI and player backlash against it, and Ubisoft's decision to test quietly inside a flagship title reflects exactly how fraught that balance has become.
Potential risks and opportunities
Risks
- If details of the AI experiments leak further before Far Cry 7 ships, Ubisoft faces organized player backlash campaigns that could suppress pre-orders, repeating the pattern seen with other studios that disclosed AI use mid-development.
- Ubisoft's stock and relations with creative unions could deteriorate if the in-house tooling is perceived as a precursor to reducing artist and writer headcount on future titles, particularly given ongoing SAG-AFTRA and game-dev labor organizing activity in 2026.
- Competitors (EA, Activision Blizzard) monitoring Ubisoft's stumble could use the negative quality signal to delay their own public AI commitments, slowing industry-wide adoption timelines and leaving Ubisoft exposed as the visible early mover with poor results.
Opportunities
- Third-party generative AI middleware vendors targeting game studios (Inworld AI, Latitude, Didimo) can use Ubisoft's reported quality struggles as a direct sales entry point, positioning their mature tooling against the risks of building in-house.
- Game engine platforms -- Unity and Epic Games -- could accelerate native generative AI feature integration as a competitive differentiator, capturing studios reluctant to invest in proprietary tooling after seeing Ubisoft's R&D friction.
- Specialist AI quality-evaluation firms and red-teaming consultancies gain a credible pitch to AAA studios for structured AI output benchmarking before experiments reach the production pipeline stage.
What we don't know yet
- Which specific capabilities are being tested inside Far Cry 7's environment -- NPC dialogue generation, environment asset creation, or procedural quest design -- has not been disclosed.
- Whether Ubisoft's in-house AI tooling is built on a licensed foundation model or trained entirely on proprietary data remains unconfirmed as of May 2026.
- How far along the Far Cry 7 production timeline these experiments began, and whether they have affected development velocity or shipping estimates, is not addressed in Henderson's reporting.
Originally reported by thegamer.com
Read the original article →Original headline: Insider Says Ubisoft Is Testing Gen AI In Far Cry 7, And It 'Looks Like S**t'