Skip to content

AI safety in the pipeline industry

Artificial intelligence (AI) is transforming industries across the board, promising increased efficiency, productivity, and cost savings.

With benefits like that, it is understandable why so many see AI as a tantalising proposition. At Klarian, we agree: AI and data analytics are core aspects of our Digipipe service. However, we refuse to allow AI to dazzle us into forgetting the potential risks. Ensuring AI safety is a critical concern, not only in the technology sector but also in industries that rely heavily on data-driven solutions, such as the pipeline industry.

If we were attending the AI Safety Summit, our stance would be that AI must be balanced with mathematics and engineering expertise. In the words of our Head of Analytics and Insight, Dr Liam Trimby.

"AI is an incredible tool for building experts out of data; but that's no excuse not to maintain our own engineering expertise. It's this expertise we lean on more than anything to intelligently train, supervise, test, and regulate our AI."

What is AI safety and why is it so important?

You wouldn’t buy a car without seatbelts.

Just like you expect cars to have safety features, so should you expect safety from AI solutions. If used improperly, AI has potentially dangerous capabilities. AI based systems can be even more vulnerable thanks to their black-box nature and unpredictability when encountering scenarios outside their training. The bottom line? Ensuring these systems are not just functional but safe and reliable is paramount. Think about self-driving vehicles that must avoid accidents or medical AI systems that aid in diagnosing diseases accurately.

Now let’s shift our focus to the pipeline sector. Pipelines carry valuable resources, so any disruptions or accidents can spell disaster, both environmentally and economically. Remember the Colonial Pipeline incident in 2021? A ransomware attack led to complete operational shutdown and fuel shortages in several states.

It is imperative that we don’t let ourselves get carried away by the possibilities of AI and forget AI safety. We need principles and practices to ensure AI technology is reliable, robust, and secure. Outputs from the AI Safety Summit will hopefully be the first step in this direction.

AI in the pipeline industry

Pipeline operations generate lots of data, and analytics and AI offer a way to harness this data’s full potential. For example, as part of our Digipipe service, we offer tailored analysis and consultancy for your system with a variety of applications, including:

  • Network optimisation - Our interoperable platform optimises the complicated relationships on a network or storage system. It reduces energy use and operational costs to provide value and RoI for customers.
  • Asset efficiency - We use a mix of AI/ML and physics-based analysis methods to evaluate performance, provide predictions and understand how assets work individually and together. These methods allow human expertise and feedback into the loop, eliminating one-then-done engineering.
  • Anomaly detection - We find and quantify unusual activity in operations data, providing early alerts for one-off problems and maintenance recommendations, resulting in lower down-time and reduced maintenance costs.
  • Predictive maintenance - We configure bespoke analytics models with our expertise in machine failure mechanisms, statistics and neural networks to improve machine health and widen maintenance response windows.
  • Virtual sensors - Guided by our industry and fluid dynamics expertise we can work with client data to build virtual sensors. These provide calibration assurance, anomaly detection, and act

AI offers pipeline operators exciting opportunities for pipeline operators to unlock hidden insights for greater efficiency and reduced emissions. As AI becomes more integrated in the operation of critical national infrastructure like pipelines, the need for AI safety becomes increasingly evident.

How Klarian ensures AI safety in Digipipe

In an era where data-driven decisions are the norm, pipeline operators cannot afford to overlook the importance of ensuring AI safety.

At Klarian, we believe we can harness the full potential of AI while safeguarding the integrity and reliability of pipeline operations by adhering to best practices and staying vigilant. In doing so, we hope to ensure a brighter and more sustainable future for the pipeline industry and the communities it serves.

We take the following steps to ensure AI safety:

  • Rigorous testing and validation processes for AI algorithms.
  • Regularly updating and maintaining AI systems to adapt to changing conditions.
  • Robust cyber security measures, such as air-gapping databases.
  • Deterministic algorithms providing oversight and sanity-checking.
  • Detach AI from direct control systems.
  • Incorporate human oversight in AI operations to address unexpected situations.

One thought lies at the heart of everything we do, we are in the protection business. We’ll be watching the outputs from the AI Safety Summit closely and recommend you do the same.