As artificial intelligence (AI) systems begin to control safety-critical infrastructure across a growing number of industries, the need to ensure safe use of AI in systems has become a top priority.
Quality assurance and risk management company DNV GL has launched a position paper to provide guidance on responsible use of AI. The paper asserts that data-driven models alone may not be sufficient to ensure safety and calling for a combination of data and causal models to mitigate risk.
Entitled AI + Safety, the position paper details the advance of AI and how such autonomous and self-learning systems are becoming more and more responsible for making safety-critical decisions.
The operation of many safety-critical systems has traditionally been automated through control theory by making decisions based on a predefined set of rules and the current state of the system. Conversely, AI tries to automatically learn reasonable rules based on previous experience.
Since major incidents in the oil and gas industry are fortunately scarce, such scenarios are not well captured by data-driven models alone as not enough failure-data is available to make such critical decisions. AI and machine-learning algorithms, which currently rely on data-driven models to predict and act upon future scenarios, may not be sufficient then to assure safe operations and protect lives.
DNV GL seeks to combine the best of the traditional physics-based methods with the opportunities provided by novel data-driven approaches.
Simen Eldevik, author of the position paper and a principle Research Scientist with DNV GL Risk & Machine Learning, said, “The emergence of AI and digital-based solutions is the next natural step for the oil and gas sector to drive efficiencies and 40% of senior oil and gas professionals say that digitalisation has improved safety over the past three years.”
“The industry is already developing and working with autonomous robots capable of performing a plethora of complex actions, including reading dials and gauges and navigating around obstacles on offshore assets.”
“However, a combination of data-driven models and the causal and physics-based knowledge of industry experts is essential when AI and machine learning are used to inform or make decisions in safety-critical systems.”
The position paper stresses that if the industry can supplement these data-driven models by generating physics-based casual data, it will be significantly closer to the safe implementation of AI in safety-critical systems.
DNV GL has joined forces with Norway’s largest universities and companies, including Equinor, Kongsberg Group and Telenor, to establish a Norwegian ‘powerhouse’ for AI. The Norwegian Open AI Lab aims to improve the quality and capacity for research, education and innovation in AI, machine learning and big data.
“We are supporting the industry to confidently make use of the most appropriate modelling and analytical approaches to better understand AI and reduce the high-risk scenarios we have pledged to safeguard against,” Liv A. Hoven, DNV GL’s CEO, added. “As the sector accelerates its digital ambitions, confidence and trust in AI will be a huge step forward in its adoption and implementation.”
A significance for the gases industry?
gasworld readers might well find themselves asking, why is this story, about the oil and gas sector, significant to us?
Well, on the face of it, this is of course not industrial gas-related. But could there be some confidence to be gleaned from this where the future role of AI and digitisation is concerned, and actually what gains there are to be made as a result of its implementation?
There is a natural reticence to place trust or confidence in AI and digitisation as a whole; it’s a step into the unknown, a game-changer, and there is always a certain sense of hesitance or anticipation around such change.
The industrial gases business is a very traditional industry, based upon strong cultural and organisational paradigms. It’s still very customer-facing – the behaviours, interactions and relationships in the industry are still very personal in nature – and above all, safety is paramount. The gases industry is built upon the Zero Ambition.
The impact of digitisation on safety and the human factor are understandably two of the biggest pain points or areas of concern for the industrial gases business, from the feedback I’ve received over the last 12 months.
There’s little doubting the productivity benefits or ROI over a relatively short timeframe, but there is confusion if not anxiety where upskilling, adaptation and job security are concerned. And there is a valid point where safety is involved, ensuring that crucial factor is not overlooked.
But if a colossal sector such as the oil and gas industry believes there are significant gains to be made – and as many as 40% of oil and gas professionals surveyed by DNV GL believe it enhances safety too – then that is surely a talking point for the industrial gas industry to take forward.
You could argue that while vastly different, the oil and gas sector is one of the most relatable industries for the industrial gases business and perhaps then, this story takes on more relevance today. The fact that autonomous and self-learning systems are becoming more and more responsible for making safety-critical decisions in the oil and gas sector, where major incidents are thankfully not a regular occurrence, is significant to our industry and its ongoing adoption of digital technologies.