Google Gemini User Trust Plummets as 65% Report Worsening Performance in Recent Poll

Growing User Dissatisfaction With Google Gemini

Google Gemini faces mounting criticism as a recent reader poll reveals declining user trust1

1

. Over 60% of more than 1,200 respondents reported that Google Gemini has gotten worse in recent months1

1

. When excluding the 5.6% who don't use the AI assistant at all, disgruntled users account for just under 65% of votes—nearly two in every three respondents1

1

. This widespread user dissatisfaction signals a troubling trend for Google's flagship AI assistant and its broader Gemini AI model family.

Reports of Degraded Performance and Reduced Response Quality

Users have flagged multiple issues affecting Gemini performance across various platforms. Reddit threads highlighted frequent misunderstandings, degraded performance, and virtually no perceptible performance differences between premium tiers1

1

. One user, Kevin Fleming, noted that responses now use less compute power, resulting in reduced response quality by a factor of approximately five1

1

. Another reader explained their switch to Claude Code, stating that Gemini is "just frustrating, and barely better than Google search results"1

1

.

Technical Issues Compound Reliability Concerns

Reliability problems have compounded declining user trust throughout 20262

2

. August brought latency issues, endless loading screens, and Error 1076 reports2

2

. October saw a content-generation outage, while February experienced a Vertex issue causing API errors for nearly two hours2

2

. On October 9, Google introduced automatic thinking-level routing, which made speed and response quality feel uneven from one prompt to the next2

2

. These technical disruptions have eroded confidence in the platform's stability for both casual and professional users.

Benchmark Scores Versus Real-World Use

Despite Google's continued updates, including the October 28, 2026 launch of Gemini 4 Carbon with a reported score of 59 on the Artificial Analysis Intelligence Index2

2

, independent analysis reveals a gap between benchmark scores and everyday use2

2

. Reported hallucination rates reached 88% for Gemini 2.5 Pro and 91% for Gemini 3 Flash2

2

. The Gemini AI model also ranked last in one source-citation comparison2

2

. These figures suggest that performance metrics don't translate to reliable real-world use, particularly for tasks requiring accuracy and source verification.

What This Means for Users and Google's AI Strategy

The reader poll results matter for anyone relying on Gemini for search, writing, studying, or routine tasks2

2

. Not everyone believes the experience is regressing—just under 20% voted that Gemini hasn't declined, while 14.6% remain unsure1

1

. However, when a large portion of users raise concerns about an AI assistant deployed across Google apps, Google Search, Android features, and API services2

2

, it signals a need for Google to address both perceived and actual performance issues. Watch for how Google responds to this declining user trust and whether future updates can restore confidence in Gemini's capabilities.