• by Laurie Sullivan , Yesterday

Google's AI Overviews and OpenAI's ChatGPT have very different ways to evaluate, characterize and criticize brands in query results -- especially as consumers near the point of purchase.

Consumers increasingly use AI not just for answers but also to evaluate brands. But the challenge emerges when AI wants to share its own opinion directly in responses. It creates a major risk to brands, especially those weak in terms of optimizing content.

AI Overviews is 44% more likely than ChatGPT to surface negative brand sentiment overall, but ChatGPT focuses its criticism 13 times more heavily near the point of purchase, when consumers are ready to decide on the products, according to new research from BrightEdge, which estimates more than three billion people now interact monthly with the two platforms.

Both engines summarize a brand’s entire digital history including reviews, forum discussions, news coverage, and past controversies, but present answers to queries, as well as conducting conversations with consumers about brands in different ways.

Negative sentiment may seem rare, especially when looking at a pie chart, but it reaches millions of consumers monthly by surfacing in approximately 2.3% of mentions of brands in AI Overviews, and about 1.6% in ChatGPT.

It's not clear how these results would affect the ads serving alongside the information. 

A 2.3% negative rate translates to millions of negative, brand mentions monthly. Unlike a buried review on page two of search results, a negative AI response is served repeatedly to every user asking a similar question, systematically redirecting demand at scale.

AI Overviews skew heavily toward controversy and negativity like lawsuits, boycotts, data breaches, regulatory actions, and product recalls. ChatGPT skews toward product-evaluation negativity like compatibility limitations, feature shortcomings and “is it worth it?” assessments.

For example, a major retailer may face negative sentiment in AI Overviews because of a newsworthy lawsuit, while results for the same retailer appearing in ChatGPT would face criticism over a specific product limitation or payment policy.

This happens because each of the AI engines pulls from different data sources. Google’s AI Overviews lean heavily into news-driven sourcing and controversy indexing. ChatGPT more frequently reflects product reviews, forums, and social discussions such as Reddit.

Negative sentiment varies by industry and engine. Apparel has an 0.2% score in AI Overviews, whereas it has 0.6% in ChatGPT.

On overlapping negative prompts, Google and ChatGPT flagged different brands 73% of the time. The data shows optimizing for one engine is not enough, according to the report.

Brand controversies and legal issue queries trigger the most negative brand sentiment, followed by product limitations, safety and recalls, service failures and outages, product discontinuation, price and value criticism, and competitive comparisons.