In recent months, there’s been a great deal of buzz and excitement around the latest advances in artificial intelligence. And while there’s no questioning this excitement, to know where you’re going, you must also remember where you’ve been. At Amazon, we’ve invested heavily in the development of AI and machine learning (ML) over the last 25 years. This should give you a good sense of how far we’ve come on the AI journey—yes, while we’ve taken the first very important steps, out on the horizon is potential for unprecedented transformation and innovation. And this is true for all industries, including oil and gas.
For the oil and gas industry specifically, those early steps into the AI journey have been impactful, and they are playing an integral role in helping the energy industry digitally transform, and address, complex business challenges like asset management and decarbonization. ML and AI have been leveraged for a wide variety of use cases, from helping to optimize artificial lift programs to emissions monitoring for pipelines and refineries.
Progressing along this AI journey, we’re now seeing the next wave of widespread adoption of ML with generative AI, a subset of ML that is powered by ultra-large models, such as large language models (LLMs) and multimodal models. Generative AI brings new capabilities, like code generation and the creation of images, based on natural language prompts, in addition to transforming existing ML-powered capabilities like enterprise search and chatbots. Generative AI is powered by foundational models, which are large ML models that have been pretrained on vast amounts of data.
When we think about the potential future of AI in the oil and gas industry, we often have sci-fi visions of completely autonomous drilling or robots managing offshore production operations. And while some of these scenarios may one day come to fruition, we’re currently just a few steps into this journey. In these early days, where we see the most potential, it is leveraging this transformative technology to address some of the oil and gas industry’s most pervasive challenges.
Knowledge management is one particular challenge, where we believe generative AI could have a transformative impact. Oil and gas organizations are often dealing with extremely large and unstructured sets that can be decades old. This results in data access and data quality issues that can have a direct impact on day-to-day activities. For example, geoscientists can spend up to 70% of their time searching for, and managing, data. These issues also can impact knowledge transfer from one generation of domain experts to the next—an issue that has been discussed across this industry for many years.
With a generative AI application, all of these data can could be ingested and integrated. Generative AI can could then extract, summarize or augment the data via natural language interactions with operators, gaining new insights from data that were buried before, or by simplifying user interactions. This could help domain experts quickly retrieve data sets and stored knowledge. And by making data discovery and access as simple as asking Alexa to help you find a recipe or driving directions, those experts would be freed up to focus on more complex challenges.
Increasing efficiencies and reducing downtime. Similar to knowledge management, the oil and gas industry is continuously looking for opportunities to increase operational efficiencies and reduce downtime. Generative AI’s synthetic data and image generation capabilities could have a profound impact in this space, as it could be a critical tool for predictive and preventative maintenance.
Synthetic data are a class of data that is generated rather than obtained from direct observations of the real world. Synthetic image creation could be used for creating images of operational equipment—pipelines, compressors, turbines, etc.—showing signs of deterioration like rust or cracks. These images could be used to train vision-based ML models that can alert workers when early deterioration signs are detected. Capturing the condition of operational equipment in advance could help companies anticipate failures before they happen. This would enable companies to better plan and optimize how, and when, they address operational issues to decrease downtime and improve overall efficiencies—not just operationally but also logistics and planning.
Enhancing subsurface knowledge and understanding is another area where the oil and gas industry has spent decades working to improve it. And now, as applications like geothermal and carbon capture and storage grow in adoption, the need to improve subsurface understanding to make decisions better and quicker will continue to increase.
Generative AI models could be used for reservoir modeling by generating synthetic reservoir models that could simulate reservoir behavior. General adversarial networks (GAN) can be used to generate synthetic reservoir models. The generator network of the GAN can be trained to produce synthetic reservoir models that are similar to real-world reservoirs, while the discriminator network is trained to distinguish between real and synthetic reservoir models. Once the generative model is trained, it could be used to generate a large number of synthetic reservoir models that could be used for reservoir simulation and optimization, helping to reduce uncertainty and improve hydrocarbon production forecasting.
And this is just the start. With cloud computing and the ready availability of scalable computing capacity, the industry is primed to leverage generative AI for hundreds of use cases. At AWS, we’re working with customers on a wide range of generative AI use cases, from energy trading compliance to report generation, while also looking at bigger challenges for safety and decarbonization.
Generative AI provides the oil and gas industry with a true roadmap to address some of its most pervasive challenges. In these early stages, the most common generative AI applications will serve as chat or natural-language interfaces that complement existing ML- and AI-based systems. But we’re just a few steps into this journey, and it’s clear that generative AI has the potential to reinvent every customer experience and application across the oil and gas value chain.
- Digital transformation: Digital twins help to make the invisible, visible in Indonesia’s energy industry (January 2024)
- Digital transformation: A breakthrough year for digitalization in the offshore sector (January 2024)
- Quantum computing and subsurface prediction (January 2024)
- Digital transformation: Building cyber resilience in the oil and gas industry (December 2023)
- Digital transformation: Putting people in the driving seat of digital uptake and scalability (December 2023)
- Digital completions platform provides complete operations visibility to enhance efficiency, collaboration (December 2023)
- Applying ultra-deep LWD resistivity technology successfully in a SAGD operation (May 2019)
- Adoption of wireless intelligent completions advances (May 2019)
- Majors double down as takeaway crunch eases (April 2019)
- What’s new in well logging and formation evaluation (April 2019)
- Qualification of a 20,000-psi subsea BOP: A collaborative approach (February 2019)
- ConocoPhillips’ Greg Leveille sees rapid trajectory of technical advancement continuing (February 2019)