Oil Reserve Estimates Move to AI as Audit Standards Lag Behind

The shift from hand-drawn curves to black-box models

The petroleum industry is quietly replacing a century-old forecasting method with machine learning, and the regulatory framework governing those forecasts hasn't kept pace. According to reporting by Forbes contributor Dara-Abasi Ita, oil companies are moving from traditional decline-curve analysis—where engineers manually plot a well's falling output and project it forward—to AI models that process more data and identify patterns standard equations miss.

Proved reserves drive most of an oil company's paper value. They feed the standardized present-value calculation (PV-10), trigger impairment writedowns, and set a five-year clock on wells a company has committed to drill. Change the reserve figure and the stock price moves with it. The difference now is that the math producing that figure is increasingly opaque.

Permian Resources' spring 2026 quarterly filing lists "artificial intelligence and its application in our industry" as a business risk, while also repeating the standard language that reserve engineering "cannot be measured in an exact way." The company's year-end 2025 reserves were prepared by outside firm Netherland Sewell & Associates, but the broader industry trend is clear: the forecast feeding the reserve number is migrating from a chart anyone can check to a model that may not show its reasoning.

Why it matters

Shareholders own a valuation partly anchored to forecasts they can't examine. Banks extending reserve-based loans—credit lines resized every six months based on a borrower's booked barrels—are lending against harder-to-verify numbers. And reserve auditors certifying these estimates are working to standards written before machine learning existed. The gap creates legal, financial, and reputational exposure across the capital structure, yet no major enforcement action or standard revision has addressed it.

What the rules actually require

SEC Regulation S-X, Rule 4-10(a)(22) defines a proved reserve as a quantity that "can be estimated with reasonable certainty to be economically producible" using deterministic or probabilistic methods. Both are forms of engineering a person can trace. The rule doesn't mention machine learning.

The SEC does permit "reliable technology"—defined as methods "field tested to provide reasonably certain results with consistency and repeatability." That language, from a 2008 modernization, was meant to let companies adopt better technology. It wasn't written for models whose output can shift when retrained on new data.

Under Regulation S-K, companies reporting material reserve additions must provide a summary of the technology used, though that summary "may be general in nature." Whether a line like "we used a proprietary machine-learning model" tells an investor enough to judge reliability remains an open question.

The audit gap

Third-party reserve auditors—firms like Ryder Scott, Netherland Sewell, and DeGolyer & MacNaughton—test the result against well data and their own judgment, not against which software produced it. If a model's forecast matches production history, by that logic the software behind it doesn't change the verdict.

But reproducing an opaque model and explaining why it weighted one input over another is a different job than re-running a decline curve. The certification standards these firms follow, and the industry's SPE-PRMS framework, were written for deterministic and probabilistic work. They don't contemplate the reproducibility or explainability of a model.

What to watch

Three markers will show whether the gap narrows. The spring reserve-report cycle, when operators file third-party exhibits, is where any change in how auditors treat model-derived forecasts would surface. SEC comment letters are the second signal—staff already ask companies to explain "reliable technologies" behind material reserve additions. The third is the reserve-auditing standards themselves: the SPE-PRMS framework is where the profession would have to write explainability into the job.

The details were first reported by Dara-Abasi Ita for Forbes.