AI, ChatGPT in oil, gas, energy and elsewhere

Accenture – ChatGPT a defining tech trend for the next decade. SLB unlocks AI potential for energy. Data Kinetic launches applied AI solutions for oil and gas. IBM Institute for Business Value on the pivot away from STEM skills. Generative AI at Mitsui Chemicals. iGenius Crystal in Microsoft Teams. Dataiku and Databricks publish insights from senior AI professionals. NIST AI Risk Management extended to LLMs. EU Centre for Algorithmic Transparency.

Stephanie Jamison, Global Resources industry leader with Accenture, in a LinkedIn Pulse post described generative AI, such as ChatGPT, as one of the defining technology trends of the next decade. Models trained on ‘internet-scale’ data are ‘incredibly powerful and infinitely adaptable’ and can be repurposed for industry use cases, in days. Potential energy usage could include optimizing efficiency at a chemical plant, dynamically balancing a diverse energy asset portfolio, transforming customer service at scale in utilities, or accelerating the well advisory process for oil and gas. Accenture has worked with a west coast utility leveraging Microsoft Azure’s GPT models to build a ‘hyper-intelligent assistant capable of swiftly processing and analyzing text, empowering employees to speed up decision-making’. The company is also ‘helping one of the largest oil and gas companies in the world manage the mountains of data it has accumulated over the decades’, ‘introducing capabilities such as multi-modal data handling, cognitive search, and semantic modelling, as well as generative AI from Microsoft Azure OpenAI’. At the human level, an oil and gas team with little experience of the LNG or hydrogen value chains might use generative AI as an expert ‘co-pilot’, enabling the team to understand customer needs and core processes. In a siloed industry such as oil and gas, ‘where each part of the business has tended to operate in its own pocket’, generative AI ‘offers the promise of stitching the value chain together end to end — from underground to overground to corporate to field operations’.

SLB recently issued a position paper on ‘Unlocking the potential of AI for the energy industry’. In oil and gas, unlocking AI’s full potential requires ‘embracing open and secure data platforms, rapid experimentation, new technology adoption and a commitment to building up an AI-trained workforce’. To operationalize AI at scale, industry must shift toward modern open data architectures based on common standards that make it easy to discover, clean, enrich, access and consume data. Enter OSDU, The Open Group Open’ Subsurface Data Universe, now ‘available worldwide’ from ‘digital technology leaders’ such as Microsoft, Google, Amazon and IBM. SLB is also offering a ‘Domain Data Scientist’ training program that sets out to ‘transform thousands of domain experts into data scientists’. SLB opines that generative AI and large language models can enable semantic search and power smart assistants to extract insights from data. Foundational models in subsurface, drilling and production domains can accelerate the automation of well construction and production operations. ‘The role of AI in the energy industry cannot be overstated’. An AI-trained workforce can harness AI’s transformative capabilities to accelerate the energy transition, paving the way for a ‘more sustainable and prosperous world’.

Austin, TX-headquartered Data Kinetic has launched a suite of Applied AI Solutions for Oil and Gas and is proposing AI transformation workshops to US-based operators who would like to explore the potential of these new solutions. ‘AI Outcomes as a Service’ for oil and gas is a platform agnostic, subscription-based catalog of machine learning models designed to accelerate enterprise adoption of applied AI. Target use cases include optimizing gas lift processes ‘at the edge’, predictive maintenance and exploration data analysis.

A publication from the IBM Institute for Business Value aggregates five years of survey data on business use of AI and LLMs. The report takes a deep dive into the thorny question of AI’s impact on jobs, making the bold claim that ‘AI won’t replace people—but people who use AI will replace people who don’t’. This means that ‘some business leaders are rushing to reorganize, elevating new skills while deprioritizing those that have become obsolete’. IBM cites a ‘study’ from the World Economic Forum that envisages the ‘disruption’ of 85 million jobs globally between 2020 and 2025 but heralds the creation of 97 million new job roles. Generative AI could push these figures even higher. The IBM IBV survey has been tracking workforce skill sets since 2016 and has observed that STEM* skills are plummeting in importance, dropping from the top spot in 2016 to 12th place in 2023. ‘As technology becomes more user-friendly, employees are also able to do more with less advanced technical skills. No-code software development platforms let people without a programming background create business critical prototypes and apps. Plus, as machines take over mundane tasks, people can spend more time on the problem-solving and collaborative work that require stronger people skills. This pivot away from STEM skills highlights the volatility of the talent landscape.’ Looking more closely at the IBV graphic we observe that the category of ‘basic computer and software application skills*’ has fallen even further, down from N°2 to N° 16’. IBM IBV research found that 83% of executives say generative AI will reinvent the way their organization works.

IBM IBV also reports on new applications for generative AI as deployed at Mitsui Chemicals. The company is working on new application discovery, leveraging a generative pre-trained transformer (GPT) in a foundation model that is now used across the Group. IBM is implementing a ‘third-party large language model’ to generate new applications from a large text corpus including ‘everything from patents to business news’. You may think that this sounds a bit like IBM Watson, and indeed, the source publication is titled ‘Combined Generative AI with IBM Watson for New Application Discovery’ (in Japanese).

* Science, technology, engineering and math.

Microsoft has teamed with iGenius to embed its ‘Crystal’ generative AI platform within Microsoft Teams. Users can query data in natural language across various fields, including financial or operational data analysis. iGenius’ Crystal has been used in oil and gas sales management to generate recommendations tailored to projects in the oil and gas oil industry. Crystal has also been deployed by Italian utility Enel to leverage conversational technology to monitor plant capacity, and temperature and maintenance levels across its global thermal generation plants. More from Microsoft.

A new report from Dataiku* and Databricks ‘Insights From 400 Senior AI Professionals on Generative AI’ finds that about 2/3 of respondents said that in the coming year, trialing LLMs in their business was likely with 45% already experimenting the technology. Unsurprisingly, executive support is a prerequisite for AI and AI ‘pioneers’ are those with a larger budget for the technology. The report plods through various AI use cases, contrasting behaviors of ‘pioneers’ and ‘the rest’, although the difference does not appear all that great. With regards to ChatGPT, 25% of the pioneers believe that it will have many valuable applications and companies should start developing those now. Some 40% of the pioneers believe that ChatGPT ‘could have valuable applications, but it’s not clear what they are’ as do 30% of ‘the rest’. Even more confusingly, 33% of all respondents are ‘unclear on what the valuable applications of ChatGPT are’. The study dips in and out of issues on data quality, on AI and LLMs which makes the overall trends rather hard to discern. However some ‘findings’ stand out: 55% of respondents are more worried than excited about the future of AI’. Here, Dataiku has created a ‘reliable, accountable, fair, and transparent’ framework to prepare enterprises for the arrival of a wave of policy proposals seeking to protect both workers and consumers from potential harms of AI. Download the RAFT Framework here. Perhaps the most puzzling chapter of the 30 page report concerns the ‘actual ROI’ that AI generates. As we read it, the oil gas and industrial sector achieves an ‘ROI per dollar spent’ in the 1-2% range. Not exactly overwhelming!

* Dataiku is an SLB partner, adding its AI functionality to the already ‘cognitive’ Delfi.

The US National Institute of Standards and Technology (NIST) is launching a new public working group on artificial intelligence (AI) that will build on the success of the NIST AI Risk Management Framework to address this rapidly advancing technology. The idea is to extend the AI RMF to generative AI. Earlier this year the National Artificial Intelligence Advisory Committee delivered its first report to the president, identifying areas of focus for the committee for the next two years. The full report, including all of its recommendations, is available on the AI.gov website. More from NIST.

In her September 2023 State of the Union, EU President Ursula von der Leyen introduced the EU Centre for Algorithmic Transparency. ECAT sets out to assure a safer, more predictable and trusted online environment for people and business. The unit will advise the Commission on governance of very large online platforms (VLOPs) and very large online search engines (VLOSEs).

Click here to comment on this article

Click here to view this article in context on a desktop

© Oil IT Journal - all rights reserved.