A new BrightEdge report found that Google’s AI Overviews are 44% more likely to show negative sentiment about brands than ChatGPT. The study analyzed hundreds of millions of prompts and searches between mid-January and February 2026.
Most AI mentions of brands were still positive or neutral, but negative responses appeared in about 2.3% of Google answers compared to 1.6% on ChatGPT — a difference that could still reach millions of users due to the scale of both platforms. BrightEdge CEO Jim Yu said one key shift with AI answers is that they often express an “opinion,” which could significantly change how marketing works online.
Google pushed back on the report, calling the methodology flawed and noting the difference in sentiment is less than one percentage point. The company also said AI Overviews simply reflect what sources on the web say about a topic.
The research also showed that negative AI responses are most often triggered by brand controversies, legal issues, product limitations, recalls, or service outages. As AI becomes a major way people research products, companies are increasingly focusing on strategies like generative engine optimization (GEO) and answer engine optimization (AEO) to influence how they appear in AI responses.