Energy Perception Index: How AI Answers Frame Oil and Gas

How AI assistants and Google frame natural gas, pipelines and the oil and gas industry.

When someone asks an AI assistant whether pipelines are safe, whether natural gas beats coal, or whether the industry is good for the country, the answer they read is the industry's reputation in that moment, written by a machine from whatever sources it found. This index scores those answers on a scale from 0 to 100, where 50 is neutral.

The October 2026 reading: 50.8 out of 100

Across ChatGPT, Gemini and Perplexity the index stands at 50.8, with a 95% range of 49.6 to 51.9. A score near 50 sounds bland. It is not: 35% of answers leaned favorable toward the industry, 33% were balanced and 32% leaned unfavorable. That middle score is two camps of roughly equal size cancelling out.

Perplexity is the only assistant whose whole range sits below 50, and the gap between it and ChatGPT is larger than the noise. Google's AI Overviews are scored separately and land at the same 50.8 as the three assistants combined.

Where the industry wins and where it loses

One overall number hides a wide spread. By topic, scores run from below 20 to above 80.

On practical questions the industry wins: what natural gas costs, what gets made from it, and the safest way to move it. It loses on climate and on what it is like to live next to the infrastructure. Climate harm came up in 41% of all answers, methane leaks in 36%, and the risk of living near a pipeline in 35%.

Search tells the same story, louder

Among Google searches that express a view, people look up the industry's risks about 13 times as often as its benefits, counted by average monthly searches.

Each channel also draws on different sources. Social and video sites hold over a third of Google's first page for these topics, yet supply almost none of the evidence inside the AI answers.

A communications plan built for one of these channels leaves the other to whoever shows up. Institutions shape the AI answer. Social and video sites shape the results page around it.

How the index is measured

We put a fixed set of everyday questions about natural gas, pipelines and the industry to ChatGPT, Gemini and Perplexity with live web search on, several times each, and to Google. Every answer is recorded and scored for whether it leans favorable, balanced or unfavorable toward the industry. Those scores are combined into one number from 0 to 100. Ranges are 95% intervals from resampling the recorded answers. The search figures come from average monthly Google search volume for phrases about the industry and from the results on Google's first page for the same topics.

What it is not. It measures how answers frame the industry, not what the public believes. It has not been checked against opinion polling, so read it as the information people walk into rather than as approval. This is the first reading, taken October 20, 2026, so it is a baseline, not a trend.

The same approach powers our Energy AI Visibility Index, which measures which energy firms AI assistants name when buyers ask who to hire.