AI and the Energy Transition

In late February I attended the Petroleum Club of Western Australia Industry Dinner on “Digitalisation, AI and Machine Learning in the Energy Sector”, billed as a discussion of how these technologies can contribute to productivity and the energy transition. The discussion got off to an early start as a participant in the Club’s Next Generation School Program described her experience in teaching students about technology using the example of a young boy given the job of tending a de-watering engine in an early English coal mine, who automated the process so that he could learn the game of marbles with his time instead 1. Her point was that automation does not threaten jobs, it gives workers the opportunity to learn new skills. The panel discussion included Anthony Brockman, General Manager of Software Integrated Solutions for Schlumberger Australasia and Far East Asia and located in Perth. His recurring theme was that Australia is a natural hub for digital leadership in the energy industry, with a robust university talent pool, access to all major energy resources (hydrocarbon, mineral and renewable), a relatively supportive government regime (for now), and a history of innovation and success (he noted that during WWII, an Australian P.O.W. led one of the most successful mass escapes from a German prison camp 2. He gave the example of a team from OMV’s Digital Excellence team in Austria who visited Perth to “see how digitalization works”, including an extended visit with Woodside, and he lamented that perhaps one of the few remaining barriers to the adoption of Artificial Intelligence is failing to see it as progress. In discussing Schlumberger’s partnership with innovative companies at their technology research lab in Palo Alto in California’s Silicon Valley, Brockman revealed that one insight obtained was that the level of collaboration was one of the only reliable predictors for success of digital initiatives, and that they would have to tackle the fear of people losing their jobs to machines in order to fully realize the potential of digital platforms like DELFI. Other panellists included Tom Georke, Innovation Centre Lead for Cisco in Perth, who observed that he considered the oil and gas industry to be only “nascent” in the adoption of machine assisted or data led decisions. He had his comparisons to the financial technology sector roundly challenged as perhaps not being the best example of using technology to maintain a “license to operate”. One of the more interesting observations of the evening came from Miranda Taylor, CEO of National Energy Resources Australia (NERA), who asserted that while many energy industries are very good at innovating internally, “you can’t scale innovation with a supply chain that operates in a silo”. This aligns well with NERA’s mandate to foster collaboration and innovation and help the energy resources sector respond to workforce trends 3. As a follow up to the discussion from Cisco, Richard Jones, VP of Asia Pacific for Dataiku, based in Singapore, polled the audience to see how many regularly used machine assisted recommendations. Only about 15% of the audience responded affirmatively, but I doubt many of them thought about letting their car GPS or public transport app tell them the best way to get to the hotel for the meeting, or odering something online that “others who ordered this also liked”. So maybe the best success criteria for artificial intelligence is that people shouldn’t realize they are using it 4. 1) https://books.google.com.au/books?id=v_-yJ5c5a98C&redir_esc=y 2) https://www.smh.com.au/world/europe/the-flight-of-the-crow-the-greatest-escape-story-never-told-20180314-p4z4e4.html 3) https://www.nera.org.au/Featured-workforce-report 4) https://arxiv.org/abs/1906.11068v1