AI SBoom or Bubble? AI-Generated Games & Films Flood Steam, Pitting Industry Insiders in an Ultimate Standoff Between Low Costs and High Quality

If you’ve browsed Steam’s new releases lately, you may have noticed a striking trend: interactive movie games are multiplying, and they share a new label — AI.

Some games use AI to generate entire video sequences, branding themselves as “cinematic FMV experiences.” Others deploy AI for character art, backgrounds, voiceovers, music, and promotional materials. Still others combine AI character chat, branching narratives, and player-created content into a hybrid format that blends “watching a show,” “playing a game,” and “chatting with AI”.

The interactive film genre, which saw a resurgence thanks to hits like Caught in the Act! I’m Surrounded by Beauties, has entered a far more complex phase. The old keywords were “live-action,” “short-drama teams,” “low-cost hits,” and “emotional value.” The new ones are “AI video,” “AI actors,” “low-barrier production,” and “UGC interactive storytelling” — along with a harsher, unavoidable critique: too much ‘AI flavor’.

The Cost Logic Has Been Upended: A Game Can Now Be Made by a Handful of People

Interactive films are nothing new. From early FMV games to Her Story, Late Shift, and the Chinese sensation Caught in the Act!, the core formula has remained consistent: using cinematic presentation to deliver choice-driven, branching narratives with multiple endings. Players make choices at key junctures, altering relationships, plot direction, and outcomes.

But live-action interactive films have a natural ceiling: production cost. Actors, locations, costumes, lighting, cinematography, post-production, reshoots — every link costs money. More importantly, interactive films don’t shoot one storyline; they shoot many. More branches mean more footage; more player choices mean greater production pressure.

AI has struck directly at this pain point.

For small teams, AI video, AI voiceovers, AI music, AI artwork, and AI-assisted scripting mean they can produce “cinematic-looking” content at a fraction of the cost. Scenes that once required a full production crew can now be generated through prompts and iterative rendering. Branches that were difficult to reshoot can theoretically be filled in by AI. A project that once needed a coordinated film crew can now yield a playable demo from just a few people — or even one person.

As a result, Steam is seeing a surge of overtly AI-generated interactive films — some even marketing “100% AI-generated video,” “AI actors,” and “AI assets” as selling points. Reports indicate that the number of Steam games publicly disclosing generative AI use rose from 7,818 in July 2025 to over 10,000 by December — roughly 8% of Steam’s total catalog. At the June 2026 Steam Next Fest, 1,694 out of 4,227 game demos — a staggering 40% — incorporated AI-generated content.

What Players Really Reject: Not AI, but the ‘Low-Cost Feel’

Whether players actually buy these games is another matter.

Steam community discussions reveal that player attitudes toward AI content are not simply “for” or “against”. More precisely, players反感 three things.

First, the “AI flavor.” The Achilles’ heel of these games isn’t individual frames — it’s continuity. Inconsistent character faces, awkward hand movements, chaotic camera language, unconvincing emotional performances, and drifting scene details. These issues might be tolerable in isolation, but within a “cinematic narrative,” they become glaring. Interactive films sell immersion; players naturally judge them by cinematic standards — Are the performances natural? Is the editing rhythm comfortable? Is the emotion consistent? Any uncanny AI glitch pulls players right out of the story.

Second, “low cost at a high price.” Players aren’t opposed to small teams using AI, nor to experimental works. But when a game is clearly assembled from AI-generated素材 yet carries a premium price tag, it’s很容易被视为 “cash grab”. AI can lower barriers, but it cannot be a遮羞布 for laziness.

Third, a lack of “human touch.” A flood of AI interactive films, criticized for homogenized art styles, rigid narratives, and absence of emotional depth, have been collectively dismissed by players as “AI garbage”. Market feedback clearly shows that players prefer live-action productions like Caught in the Act!.

The data backs this up. A study of nearly 10,000 Steam games found that games disclosing AI use averaged 52.6% fewer first-month reviews than non-AI games. Since Steam review counts are widely considered a proxy for sales, this suggests AI-labeled games may sell roughly half as well. Even among games with at least 100 reviews, AI games had a median positive rating of 84.6%, versus 88.3% for non-AI games. Separate surveys indicate that 63% of players hold a “very negative” view of AI in games.

Technical Hurdles and Copyright Concerns: Two Shackles on AI Interactive Films

On the technical front, the coherence of AI-generated video in long-form narrative contexts remains an unsolved problem. Tencent’s AI video creation tool TDream claims to support “AI real-time generation of playable content,” but its sample film Shan Yue still suffers from unnatural transitions and unstable character proportions — suggesting that long-term continuity has yet to be fully resolved. Interactive films must simultaneously solve three core challenges: AI video generation, branching narrative logic, and user interaction responsiveness — a complexity far beyond that of short video generation.

Even more concerning is narrative logic. AI can generate 100 branching endings, but struggles to ensure that every branch aligns with character personalities and causal chains. Industry analysis suggests that the overall loss rate for interactive films is as high as 90% — AI has lowered the production barrier, but it hasn’t solved the fundamental problem of “how to make good content.”

Copyright issues loom just as large. For IP holders, character images entering UGC-driven AI interactive film platforms risk being uncontrollably “refined,” spliced, or misappropriated. Legal experts note that AI-native games must simultaneously comply with both internet information service algorithm registration and generative AI security assessment requirements. As AI evolves from tool to native game asset, compliance and copyright will become regulatory minefields that developers must navigate proactively.

Where Is the Industry Headed? The Hype Isn’t Here Yet, but the Shakeout Has Begun

Despite the controversies, industry giants haven’t slowed down. In May 2026, Tencent launched “Code-named Craft,” an AI game creation platform that generates 2D/3D games from natural language inputs, integrating complete AIGC tools and over 10,000 art assets. ByteDance rolled out Seedance 2.0, a video generation model hailed by industry media as an “all-around AI director”. In May 2026, Tencent’s Photon Studios also unveiled “Zaohua Gongfang,” an AI interactive narrative creation platform that has already accumulated nearly 60 community works. A six-person startup, powered by AI, reportedly “manages” a production team equivalent to nearly 1,000 people, releasing the world’s first AIGC-powered interactive film.

Meanwhile, live-action interactive films continue to prove their commercial worth. Shengshi Tianxia sold 150 million yuan across two installments; Emotional Anti-Fraud Simulator ranked among Steam’s top Chinese games with 1.1 million units sold. Projections indicate that China’s interactive film market is expected to exceed 1.8 billion yuan in 2026, up 20% year-over-year.

Industry players are exploring a collaborative model where “humans guard the story’s core while AI expands the boundaries of expression”. According to a CITIC Securities report, AI is empowering interactive films on two fronts — cost reduction and interactive storytelling — and future AI-native interactive films and world-model games could further unlock the ceiling of interactivity.

But when over 40% of Steam Next Fest demos carry an AI label, when “too much AI flavor” becomes the most frequent complaint in player communities, and when AI-labeled games see their sales cut roughly in half — the industry may need to pause and ask: AI has lowered the barrier, but what’s flooding in — more creativity, or more mediocrity?

Whether the next stop for interactive films is a genuine breakthrough or a passing bubble may depend less on the AI technology itself and more on whether creators can find the true balance between “low cost” and “high quality.