Google Gemini v1.2026.3370502 adds aggressive content filters, sparking user backlash in August 2026

The update from version 1.2026.3270601 to 1.2026.3370502 arrived without specific documentation from Google. The official App Store listing offers only a generic note about providing the "latest advances from our models."

The conclusion that stricter content moderation was the primary change is based on the immediate and consistent shift in user complaints after the update, rather than any official announcement.

The core of the user backlash centers on the AI's new refusal to engage with a wide range of prompts, particularly those involving creativeGlossaryAd CreativeThe visual and audio content of an ad — video, image, copy, sound. The single most important variable in modern mobile UA, lifting IPM and lowering effective CPI. writing or role-playing. A 1-star review on version 1.2026.3370502 states, "If you’re looking for roleplay, or talking about zombies or ANYTHING AT ALL related to anything over PG it wi^[2]ll just keep repeating that it cannot fulfill this request."

The same reviewer notes that mentioning "blood or anything about the undead" prompts the AI to suggest a crisis helpline. Another 1-star review from the same period calls the AI "dumb" and claims it "lacks common sense it has terrible memory."

The sentiment shift is stark when comparing reviews before and after the update. Before v1.2026.3370502, users praised the app, with one 5-star review calling it their "go to AI App to use to help organize my thoughts."

After the August 26 release, the tone soured. The average rating for reviews of the new version, v1.2026.3370502, is 3.86 stars, a fall from the 4.66-star average for reviews of the prior version, v1.2026.3270601. This pattern mirrors a similar backlash event in May 2026, when an earlier update also tightened safety filters, suggesting a recurring difficulty for Google in balancing safety with utility.A thread on Reddit from August 2026 reflects the App Store complaints, with one user stating, "It’s been lobotomized.^[3] It refuses to do anything creative."

While not officially confirmed, the stricter filters appear to be a deliberate, if poorly calibrated, move by Google. The company is likely attempting to mitigate risks associated with AI-generated content and ensure brand safetyGlossaryBrand SafetyThe controls that keep ads from appearing next to harmful or off-brand content — protecting advertisers from reputational damage and publishers from losing premium d.... However, this over-correction has damaged the app's core utility for a vocal user segment.

With over 10.9 million US iOS downloads in the last 30 days, sustained user dissatisfaction could pushGlossaryPush NotificationA message delivered to a user's device from your app's server — the single biggest re-engagement lever in mobile, when done well, and a fast path to uninstall when d... creative and power users toward competing AI platforms that offer greater flexibility.

This incident highlights a central challenge for developers of large language models: finding a workable balance between freedom of expression and content safety. Google's attempt to de-risk Gemini appears to have impaired its functionality for creative applications, a key use case for many users.

Unless Google recalibrates these filters in a future update, it risks alienating a significant part of its user base. The event serves as a case study in how AI safety measures, when implemented too broadly, can themselves become a critical product flaw. These figures compare the 1514 US App Store reviews tagged with v1.2026.3270601 against the 80 tagged with v1.2026.3370502, measured on 2026-09-05.