2 min read
Industry Contents exists because most AI coverage is either vendor marketing with a byline or hype with no numbers behind it. These are the rules we hold ourselves to, published so you can hold us to them as well.
Every claim traces to a primary source
When we cite a figure, we link to where it originated: the peer reviewed paper, the regulatory filing, the company engineering post, the court document. We do not cite an article that cites an article. If a number cannot be traced to its origin, it does not run.
Vendor data is labelled as vendor data
Benchmarks commissioned by the company selling the product can still be useful, but they are not independent evidence. Where we use them, we say who paid for the test and we say it in the same sentence as the result, not in a footnote.
We report the limits, not just the capability
An average accuracy score hides where a system fails. Wherever the underlying data allows it, we report performance by category rather than as a single blended number, because the weak categories are where the risk actually sits for the person making the decision.
How we use AI, and how we do not
We use AI tools for research assistance, drafting support and editing, in the same way any modern newsroom uses software. We do not publish unverified AI output. Every factual claim, figure and quotation in a published piece is checked against its primary source by a human before it goes live. Any error that survives that process is ours, not the tool’s.
Corrections
If we get something wrong, we fix it and we say so. Corrections are noted at the foot of the article with the date and what changed, rather than edited away silently. If you have spotted an error, tell us through the contact page and we will look at it the same week.
