CoChat
Sign Up
Cited by 285Artificial Intelligence ApplicationsImpact of AI and Big Data on Business and SocietyOil and Gas Production Techniques

Artificial intelligence in oil and gas upstream: Trends, challenges, and scenarios for the future

Open access
Koroteev, Dmitry, Tekic, Zeljko·2021·Energy and AI
DOI 10.1016/j.egyai.2020.100041
Open this paper in CoChat — free accountOpen PDF

Abstract

We analyze how artificial intelligence changes a significant part of the energy sector, the oil and gas industry. We focus on the upstream segment as the most capital-intensive part of oil and gas and the segment of enormous uncertainties to tackle. Basing on the analysis of AI application possibilities and the review of existing applications, we outline the most recent trends in developing AI-based tools and identify their effects on accelerating and de-risking processes in the industry. We investigate AI approaches and algorithms, as well as the role and availability of data in the segment. Further, we discuss the main non-technical challenges that prevent the intensive application of artificial intelligence in the oil and gas industry, related to data, people, and new forms of collaboration. We also outline three possible scenarios of how artificial intelligence will develop in the oil and gas industry and how it may change it in the future (in 5, 10, and 20 years).

AI snapshot

Written from the paper’s full text.

TL;DR

This review analyzes how AI is being applied in oil and gas upstream to accelerate processes and reduce risk, drawing on the author’s involvement in dozens of AI projects. It also outlines non-technical barriers around data, people, and collaboration, and presents positive, realistic, and negative scenarios for AI penetration over 5, 10, and 20 years.

Key findings

  • In petrophysical interpretation, an ML algorithm trained on historical well logs achieved 92% accuracy versus manual interpretation and was about 1000 times faster; a second manual interpretation by the same experts matched the first at 91% accuracy.
  • Reported or expected accelerations include seismic interpretation by a factor of 10–1000, reservoir simulation by 100–1000 times (Table 1: 200–2000), and well-treatment effect estimation 100+ times faster; AI-aided drilling is expected to reduce non-productive time by 20–40% and failures by up to 90%.
  • Scenario forecasts for AI in upstream: positive scenario predicts 40–50% cost savings at E&P by 10 years; realistic scenario predicts 10–15% cost savings; negative scenario predicts no significant margin growth and possible end of the oil century in 20–30 years.
  • Data and collaboration barriers are central: the UK’s oil and gas National Data Repository contains 130 TB covering more than 12,500 wellbores, 5000 seismic surveys, and 3000 pipelines, while U.S. AI-related companies raised $18.5 billion in 2019.

Methods

  • The paper is a review and analysis based on learnings from dozens of AI projects performed with the author’s involvement over the last three years for leading oil and gas upstream companies.
  • Projects covered geological assessment of reservoirs, drilling optimization, reservoir engineering/field development, and production optimization, as summarized in Table 1.
  • The paper reviews AI application possibilities, algorithms, data availability, non-technical challenges, and develops three qualitative scenarios for AI development in the industry over 5, 10, and 20 years.

Limitations

  • The paper does not include a dedicated limitations section. It states that its scenarios assume no major technological breakthroughs in energy and no social storms affecting energy demand trends, and that the Covid-19 crisis will not permanently change the main postulates of the industry.
  • Evident from the text: the scenarios are qualitative forecasts rather than validated quantitative models, and the evidence draws heavily on the author’s own project experience and selected industry examples.
  • Evident from the text: the paper identifies persistent data-quality, labeled-data scarcity, AI talent shortage, and limited cross-company data sharing as barriers, which constrain the generalizability and near-term applicability of its projections.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1016/j.egyai.2020.100041

Cites

48 of 48 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

0 of 56 cited sources checked · 0 supported · 0 flagged

  • Findingown data

    Internal trials for industrial partners demonstrate that a solid portion of well logging measurements could be easily reconstructed with ML.

    “internal trials we did for industrial partners demonstrate that a solid portion of the well logging measurements could be easily reconstructed with ML.”
    Quote in the paperSubstantial claim
  • Findingown data

    In the authors' automated interpretation algorithm based on machine learning, the accuracy of ML interpretation versus manual interpretation was 92%.

    “The accuracy of ML interpretation versus manual interpretation was 92%.”
    Quote in the paperSubstantial claim
  • Findingown data

    The ML interpretation was about 1000 times faster than the manual interpretation.

    “The ML interpretation was about 1000 times faster than the manual.”
    Quote in the paperSubstantial claim
  • Findingown data

    A second manual interpretation of the same data by the same experts versus the first manual interpretation showed an accuracy of 91%.

    “the second manual interpretation versus the first manual interpretation showed an accuracy of 91%.”
    Quote in the paperSubstantial claim
  • Findingown data

    A tool for automated mapping of reservoir rock properties over an oil region (using none gradient optimization + interpolation techniques) speeded up the manual mapping procedure from several weeks to several seconds.

    “Tool for automated mapping | None gradient | Speeded up the | |”
    Quote in the paperNot a substantial claim
  • May combine two claims. 34
  • The quote only partly states this claim. 3
  • Classified as probably not a substantial claim. 41

  • May combine two claims.

    Findingown data

    Based on the authors' evidence, the critical challenges to AI implementation in oil and gas are related to the new profile of people the industry requires, the central importance of data, and the need for open collaboration.

    “Based on our evidence, the critical challenges are related to the (new) profile of people the industry requires, the central importance of data, and the need for open collaboration.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingnone given

    Petrophysical interpretation is a rather time-consuming process, and the result depends strongly on the interpreter.

    “The petrophysical interpretation is a rather time-consuming process, and the result of the interpretation depends strongly on the interpreter (i.e., expert).”
    Quote in the paperSubstantial claim

  • The quote only partly states this claim.

    Findingown data

    A tool for detecting drilled rock type and potential failure using real-time drilling telemetry (using a combination of machine learning algorithms) achieved up to 20% time saving and up to 15% money savings at well construction.

    “Tool for detecting the drilled | Combination of | Up to 20% time saving | Maximizing the contact |”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Findingnone given

    The well treatment forecast tool gave up to 20% growth of marginality of the investments to the campaigns.

    “Up to 20% growth of |”
    Quote in the paperNot a substantial claim

  • The quote only partly states this claim.

    Findingnone given

    A tool for accelerating conventional reservoir simulations (using deep neural networks) accelerated by a factor of 200 to 2000.

    “Tool for accelerating the | Deep neural networks | Accelerating by a | Making it possible to screen |”
    Quote in the paperSubstantial claim

  • The quote only partly states this claim.

    Findingnone given

    The reservoir simulation tool made it possible to screen through much more field development scenarios for selecting the most optimal one.

    “Making it possible to screen |”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The automated mapping tool removed human errors causing wrong mapping, making a more accurate definition of right hydrocarbon targets.

    “Removing human errors |”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Interpretationreasoning

    AI is bringing a new approach in developing oil and gas fields, one in which data is key.

    “AI is bringing a new approach in developing the oil and gas fields, one in which data is key.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    The first steps in using AI and ML in the upstream are being made and are becoming increasingly popular.

    “the first steps in using artificial intelligence and machine learning in the upstream are made, becoming increasingly popular [13] .”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [13] Shafiee, 2019, Decision support methods and applications in the upstream oil and gas sector, J Pet Sci Eng source

      Not checked in the free run.

  • Classified as probably not a substantial claim. May combine two claims.

    Premisenone given

    Companies in the upstream deal with enormous uncertainties handled manually and relied on expert knowledge, not actual data.

    “sector deal with enormous uncertainties handled manually and relied on expert knowledge, not the actual data.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Premisecitation

    Pioneering efforts on predicting the efficiency of hydraulic fracturing jobs and ML-based analysis of injectivity issues have already been made.

    “Pioneering efforts on predicting the efficiency of hydraulic fracturing jobs [40] and ML-based analysis of injectivity issues [41] have already been made.”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [40] Makhotin, 2019, Gradient boosting to boost the efficiency of hydraulic fracturing, J Pet Explor Prod Technol source

      Not checked in the free run.

    • cites [41] Orlov, 2019, Advanced analytics of self-colmatation in terrigenous oil reservoirs, J Pet Sci Eng source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Interpretationreasoning

    AI systems, trained with the right field data, can offer significant help by speeding up the asset assessment process and making it more objective or expert-independent.

    “For both questions, AI systems, trained with the right field data, can offer significant help by speeding up the asset assessment process and making it more objective or expert-independent.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    The paper's analysis utilizes learnings from dozens of AI projects performed with the authors' involvement over the last three years for leading oil and gas upstream companies worldwide.

    “The paper utilizes learnings from dozens of AI projects performed with the authors' involvement over the last three years for leading oil and gas upstream companies worldwide.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationnone given

    The success of artificial intelligence critically depends on human intelligence.

    “The success of artificial intelligence critically depends on human intelligence.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

    Interpretationnone given

    With the development and implementation of AI-aided drilling support systems working with real-time drilling telemetry, non-productive time can be significantly reduced down to 20-40% on average and failures decreased down to 90%.

    “We expect a significant reduction of the non-productive time down to 20-40% on average with a considerable decrease in failures down to 90% with the development and implementation of AI-aided drilling support systems working with real-time drilling telemetry.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Interpretationreasoning

    Strong dependence on expert opinion and the insufficiency of appropriate input data for traditional modeling tools result in biased and uncertain answers.

    “Strong dependence on expert opinion and the insufficiency of appropriate input data for the traditional modeling tools result in biased and uncertain answers.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Interpretationreasoning

    Many reservoir engineers use tricks to perform upscaling in a way that seems correct to themselves, introducing a strong bias to the reservoir model.

    “many reservoir engineers use tricks to perform it in a way that seems correct to themselves.It introduces a strong bias to the reservoir model.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    Further development of algorithms based on optimization math and programming will enable full-scale recommending systems for well treatment design and campaign planning.

    “We expect that further development of algorithms based on optimization math and programming will enable full-scale recommending systems.The recommending systems will help select the particular well treatment design for a particular well and plan the well treatment campaigns.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    To actively use AI in processes and products, companies must grow in-house teams composed of data and AI specialists.

    “to actively use AI in processes and products, companies must grow in-house teams composed of data and AI specialists.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S1 (asserted), S2 (checkable), S3 (load-bearing), S4 (based). May combine two claims.

    Interpretationnone given

    Machine or deep learning could make history matching faster and less biased.

    “trying to involve a machine or deep learning to make history matching faster and less biased.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Interpretationnone given

    There is a low probability that AI techniques will optimize the physical part of the first seismic surveying at an asset.

    “There is a low probability that the AI techniques will optimize the physical part (i.e., amount, cost, and placement layout of sensors) of the first seismic surveying at an asset.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based). May combine two claims.

    Interpretationnone given

    AI-tools exclude people from many processes and lead to numerous possibilities for operational and business model innovations, making it possible to do things differently at the architectural level.

    “AI-tools exclude people from many processes and lead to numerous possibilities for operational and business model innovations, making it possible to do things differently at the architectural level.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    The upstream is the most capital-intensive and important of the three segments in the oil and gas business.

    “The upstream is of particular interest as it is the most capital-intensive and important of the three segments in the oil and gas business [13] .”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [13] Shafiee, 2019, Decision support methods and applications in the upstream oil and gas sector, J Pet Sci Eng source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    AI solutions are not generic and cannot be just bought.

    “AI solutions are not generic -they cannot be just bought.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [46] Ng, 2016, What artificial intelligence can and can’t do right now, Harv Bus Rev

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Interpretationreasoning

    A deep learning algorithm trained on multiple cases of manual upscaling could increase objectiveness and speed of the upscaling process.

    “This could be done well with a deep learning algorithm trained on multiple cases of manual upscaling.The outcome here would not be only the increased objectiveness but also the increased speed of the upscaling process.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    The primary target of AI and other digitalization efforts in oil and gas companies is to improve efficiency, typically by accelerating processes and reducing risks.

    “their AI's primary target (and other digitalization) efforts are to improve efficiency.In practice, that typically means to accelerate processes and reduce risks [ 8 , 11 , 13 ].”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites [8] Li, 2020, Applications of artificial intelligence in oil and gas development, Arch Comput Methods Eng

      Not checked in the free run.

    • cites [11] Lu, 2019, Oil and gas 4.0 era: a systematic review and outlook, Comput Ind source

      Not checked in the free run.

    • cites [13] Shafiee, 2019, Decision support methods and applications in the upstream oil and gas sector, J Pet Sci Eng source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationreasoning

    Generative architectures of deep neural networks, if properly trained on multiple manually conducted extrapolation exercises, can accelerate the process by a factor of 1000+.

    “we expect that, if properly trained on multiple manually conducted extrapolation exercises, generative architectures of deep neural networks can accelerate the process by a factor of 1000 + .”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: The interpolation of petrophysical properties into the seismic cube is the most time-consuming and subjective part of the geology modeling workflow.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Goodfellow, Bengio, and Courville estimated that around 5000 labeled examples are needed for acceptable performance with supervised deep-learning, and at least 10 million labeled samples are required to match or exceed human-level performance.

    “Goodfellow, Bengio, and Courville [47] estimated that to achieve somehow acceptable performance levels with the most interesting form of today's AI (i.e., supervised deep-learning), around 5000 labeled examples are needed for training.While to match or exceed human-level performance, at least 10 million labeled samples are required.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [47] Goodfellow, 2016

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning

    The development of these greenfields requires expensive technologies and makes the profitability of further oil production questionable.

    “The development of these greenfields requires expensive technologies and makes the profitability of further oil production questionable.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Premisecitation

    There are known issues with the quality and accuracy of field data and overall lack of large volumes of labeled data in the oil and gas industry.

    “Still, it is not a guaranty for success as there are known issues with the quality and accuracy of field data and overall lack of large volumes of labeled data in the oil and gas industry [48] .”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [48] Hajizadeh, 2019, Machine learning in oil and gas; a SWOT analysis approach, J Pet Sci Eng source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    The oil and gas industry is famous for its risk-averse culture and poor innovation management practices.

    “This may be one of the core problems for the oil and gas industry famous for its risk-averse culture and poor innovation management practices [48] .”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [48] Hajizadeh, 2019, Machine learning in oil and gas; a SWOT analysis approach, J Pet Sci Eng source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    BP invested in Belmont Technology, a startup aiming to boost the company's AI and digital capabilities in its upstream offshore business.

    “BP invested in Belmont Technology, a startup aiming to boost the company's AI and digital capabilities in its upstream offshore business [68] .”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [68] offshore-technology.com, “BP invests in artificial intelligence for upstream business,” 2019. https://www.offshore-technology.com/news/bp-artificial-intelligence-investment/ (Accessed 23 Apr 2020).

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    The UK's oil and gas National Data Repository contains 130 terabytes of geophysical, infrastructure, field, and well data, covering more than 12,500 wellbores, 5000 seismic surveys, and 3000 pipelines.

    “It contains 130 terabytes of geophysical, infrastructure, field, and well data, covering more than 12,500 wellbores, 5000 seismic surveys, and 3000 pipelines [64] .”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [64] Oil and Gas Authority, “Oil and gas authority open data,” 2019. https://data-ogauthority.opendata.arcgis.com/ (Accessed 23 Apr 2020).

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    This created culture of free sharing and open publishing, which companies across industries had to embrace as a standard to succeed in the era of AI.

    “This created culture of free sharing (e.g.GitHub) and open publishing (e.g.arXiv), which companies across industries (and across the globe) had to embrace as a standard to succeed in the era of AI [ 62 , 63 ] once they joined the race.”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [62] Chesbrough, 2003, The era of open innovation, MIT Sloan Manag Rev

      Not checked in the free run.

    • cites [63] Von Hippel, 2016

      Not checked in the free run.

    • Rests on an assumption: AI is born in open and collaborative environment as a consequence of academia being a leading force in AI research for decades, almost without any business influences.

  • Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

    Interpretationreasoning

    The adoption of AI demands not just data science and AI skills but a new way of thinking about problems rooted in deep understanding of the processes and core logic of tasks.

    “So, it is not that just data science and AI skills are in demand due to the adoption of AI, but a new way of thinking about problems oil and gas companies face, rooted in deep understanding of the processes and the core logic of tasks.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationreasoning

    Access to big and quality data is a crucial enabler and barrier for AI applications' successful development.

    “Thus, access to big and quality data is a crucial enabler and barrier for AI applications' successful development.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationreasoning

    Over time, the role of petroleum engineers will become crucial, as otherwise wrong questions may be asked and existing human mistakes amplified.

    “Over time, this will become a crucial role, as otherwise the wrong questions may be asked and existing human mistakes amplified”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [53] Wired, “Artificial intelligence makes bad medicine even worse,” 2020. https://www.wired.com/story/artificial-intelligence-makes-bad-medicine-even-worse/ (Accessed 23 Apr 2020).

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Ambyint has spent over a decade to build its repositories of high-quality production and optimization data that is used to train and improve their solutions.

    “One of good examples is Ambyint, a VC-funded AI-driven startup focused on oil and gas production optimization, which has spent over a decade to build its repositories of high-quality production and optimization data that is used to train and improve their solutions [58] .”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [58] GE Ventures, “AI Startup to Boost Oil & Gas Production Optimization,” 2018. https://www.ge.com/ventures/news/artificial-intelligence-startup-ambyint-announces-115-million-series-investment-boost-oil-gas (Accessed 23 Apr 2020).

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Interpretationreasoning

    The need to acquire the latest AI technology and talent are additional reasons for oil and gas companies to adopt open collaboration.

    “Next to data access, the need to acquire the latest AI technology and talent are additional reasons for oil and gas companies to adopt open collaboration.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The ability to question everything and efficiently experiment with data allows AI-born startups to attract attention and record investments.

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Interpretationreasoning

    Capability to update the pre-trained model on-the-fly will significantly increase the applicability envelope for the AI-aided tools.

    “Capability to update the pre-trained model on-the-fly will significantly increase the applicability envelope for the AI-aided tools.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    GE and Statoil jointly invested in Ambyint.

    “For example, GE and Statoil jointly invested in Ambyint [58] .”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [58] GE Ventures, “AI Startup to Boost Oil & Gas Production Optimization,” 2018. https://www.ge.com/ventures/news/artificial-intelligence-startup-ambyint-announces-115-million-series-investment-boost-oil-gas (Accessed 23 Apr 2020).

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    In 2019 AI-related companies in the U.S. raised $18.5 billion, almost 2 billion more than in 2018.

    “In 2019 AI-related companies in the U.S. raised $18.5 billion, almost 2 billion more than in 2018 [66] .”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [66] venturebeat.com, “AI startups raised $18.5 billion in 2019, setting new funding record,” 2020. https://venturebeat.com/2020/01/14/ai-startups-raised-18-5-billion-in-2019-setting-new-funding-record/ (Accessed 23 Apr 2020).

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    The UK's oil and gas National Data Repository is one of the first large oil and gas open data releases.

    “The UK's oil and gas National Data Repository is one of the first large oil and gas open data releases.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [64] Oil and Gas Authority, “Oil and gas authority open data,” 2019. https://data-ogauthority.opendata.arcgis.com/ (Accessed 23 Apr 2020).

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Interpretationnone given

    The AI-related oil and gas ecosystem comprises many different players - companies and organizations from different sectors, with different AI development stages, different strategies, and priorities.

    “The AI-related oil and gas ecosystem comprises many different players -companies and organizations from different sectors, with different AI development stages, different strategies, and priorities.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Quick adaptation studies are under intensive research.

    “Such quick adaptation studies are also under intensive research [ 60 , 61 ].”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [60] D. J. Rezende, S. Mohamed, I. Danihelka, K. Gregor, and D. Wierstra, “One-Shot generalization in deep generative models,” 2016.

      Not checked in the free run.

    • cites [61] P. Wang, L. Liu, C. Shen, Z. Huang, A. Van Den Hengel, and H. T. Shen, “Multi-attention network for one shot learning,” 2017, doi: 10.1109/CVPR.2017.658. source

      Not checked in the free run.

  • Classified as probably not a substantial claim. May combine two claims.

    Interpretationreasoning

    AI initiatives will not fail because of bad algorithms, but rather because of lack of vision, late or even no changes in the organization's operational and business model, due to lack of high-resolution data and poor collaboration.

    “AI initiatives will not fail because of bad algorithms, but rather because of lack of vision, late or even no changes in the organization's operational and business model, due to lack of high-resolution data and poor collaboration.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationreasoning

    The role and required skillset of petroleum engineers will change.

    “Just their role and required skillset will change.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Premisecitation

    Shell, Saudi Aramco, and Chevron invested together in AI startup Maana, which partnered with Microsoft to use its cloud computing platform Azure.

    “Shell, Saudi Aramco, and Chevron invested together in AI startup Maana, which partnered with Microsoft to use its cloud computing platform Azure [69] .”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [69] Houston Chronicle, “Backed by Big Oil, tech startup launches new partnership with Microsoft,” 2019. https://www.chron.com/business/energy/article/Backed-by-Big-Oil-A-I-startup-launches-new-13682135.php (Accessed 23 Apr 2020).

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable).

    Premisecitation

    Succeeding in the digital competition is not about technology only.

    “Succeeding in the digital competition is not about technology only [12] .”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [12] Tekic, 2019, From disruptively digital to proudly analog: a holistic typology of digital transformation strategies, Bus Horiz source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    There is a significant shortage of AI talent on the job market, and prospects are not good for the next decade.

    “There is a significant shortage of AI talent on the job market [ 50 , 51 ], and with more and more companies getting into AI and forming their own AI groups, prospects are not good for the next decade.”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [50] 2017

      Not checked in the free run.

    • cites [51] 2019

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    While open innovation is becoming standard in the tech sector, oil and gas companies are not famous for their joint industry projects, especially between competitors and especially not in strategic domains such as AI.

    “While open innovation is becoming standard in the tech sector, oil and gas companies are not famous for their joint industry projects, especially between competitors and especially not in strategic domains such as AI [48] .”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [48] Hajizadeh, 2019, Machine learning in oil and gas; a SWOT analysis approach, J Pet Sci Eng source

      Not checked in the free run.

    • Rests on an assumption: Oil and gas companies are not known for their agile, lean, and bottom-up development approaches but for strict jurisdiction division and waterfall processes and procedures.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Small data learning enables training the AI algorithms with a small number of examples.

    “One of the key directions here is small data learning [59] that enables training the AI algorithms with a small number of examples.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [59] M. Olson, A. J. Wyner, and R. Berk, “Modern neural networks generalize on small data sets,” 2018.

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Premisereasoning

    Some universities like Skolkovo Institute of Science and Technology and West Virginia University already started implementing special educational programs that are a healthy mix of data science and petroleum studies.

    “some universities like Skolkovo Institute of Science and Technology (Russia) and West Virginia University (US), already started implementing special educational programs that are a healthy mix of data science and petroleum studies.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    To successfully innovate in the AI-era, oil and gas companies will need petroleum engineers with a strong sense of data science and the ability to identify and design tasks to be solved by AI.

    “To successfully innovate in the AI-era, next to data scientists oil and gas companies will need petroleum engineers with a strong sense of data science and the ability to identify and design tasks to be solved by AI.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Strategy plays a key role and is a driving force of digital transformation, and top management commitment is essential to assure the success of AI and other transformative efforts.

    “Thus, strategy plays a key role and is a driving force of digital transformation [7] , and top management commitment is essential to assure the success of AI and other transformative efforts [74] .”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [7] Kane, 2015, Strategy, not technology, drives digital transformation, MIT Sloan Manag Rev

      Not checked in the free run.

    • cites [74] Roberts, 1988, What we’ve learned: Managing invention and innovation, Res Manag

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based). May combine two claims.

    Interpretationreasoning

    Data's central role and growing convergence will drive the nature of connections between members of the evolving and enlarging oil and gas ecosystem, defining how risk should be managed, value distributed and collaboration orchestrated.

    “Data's central role and growing convergance will drive the nature of connections between members of the evolving and enlarging oil and gas ecosystem, defining how risk should be managed, value distributed and collaboration orchestrated.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

    Interpretationreasoning

    The role of petroleum engineers will be to ensure that the right problems are identified for applying AI, that the right data is collected, and that solutions fit the physical and process reality.

    “Their role will be to ensure that the right problems are identified for applying AI, that the right data is collected and that solutions fit the physical and process reality.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data

    Artificial intelligence is an emerging trend in oil and gas, and there are applications that have already brought countable value.

    “It is clear that even though artificial intelligence is an emerging trend in oil and gas, there are applications that have already brought countable value.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Interpretationreasoning

    To compete with tech giants, leading universities and startups over the same talent, oil and gas companies have to fight negative attitudes toward fossil fuel industries.

    “Next, to compete with tech giants like Google, Yandex, IBM, and Amazon, leading universities and cool startups worldwide over the same talent -oil and gas companies have to fight negative attitudes toward fossil fuel industries.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Oil and gas companies have to fight negative attitudes toward fossil fuel industries to compete for AI talent.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    In April 2020, oil prices went negative.

    “The flavor of failing that industry experienced two times in the last five years, especially the last one (April 2020) in which oil prices went negative, maybe the right trigger to start transforming their business models [75] .”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [75] Andriole, 2017, Five myths about digital transformation, MIT Sloan Manag Rev

      Not checked in the free run.

60 of 60

  • The paper relies on this without support.

    In many cases, money invested in extra drilling, well treatment, or enhanced oil recovery in brownfields does not pay off.

    “In many cases, money invested in these actions does not pay off, leaving the brownfields in a slow process of dying.”

  • The paper relies on this without support.

    The interpolation of petrophysical properties into the seismic cube is the most time-consuming and subjective part of the geology modeling workflow.

    “This interpolation is probably the most time consuming and subjective part of the whole geology modeling workflow.”

  • The paper relies on this without support.

    Well treatment costs are significant and comparable with the cost of well construction.

    “The well treatment costs are significant and comparable with the cost of well construction.”

  • The paper relies on this without support.

    The amount of oil and gas recoverable from conventional brownfields with cheap waterflooding is very small.

    “the amount of oil and gas recoverable with cheap waterflooding is very small.”

  • The paper relies on this without support.

    Oil and gas companies have to fight negative attitudes toward fossil fuel industries to compete for AI talent.

    “oil and gas companies have to fight negative attitudes toward fossil fuel industries.”

  • The authors state this as an assumption.

    The absence of major technological breakthroughs within the energy area and social storms affecting current energy demand trends in the coming years.

    “Assuming the absence of the major technological breakthroughs within the energy area and social storms affecting current energy de-mand trends in the coming years, we can visualize three possible scenarios of AI spreading in the oil and gas upstream.”

  • The authors state this as an assumption.

    The Covid-19 crisis will not permanently change the main postulates of the industry.

    “Here we assume that Covid-19 crisis will not permanently change the main postulates of the industry.”

  • The authors state this as an assumption.

    Strong and committed leadership in the key companies and good data platforms are in place.

    “Assuming strong and committed leadership in the key companies and that good data platforms are in place, one may expect very rapid growth of AI capabilities for the upstream applications followed by AI tools for decision making at various levels.”

  • The paper relies on this without support.

    The algorithms and their developers are not responsible, but the responsibility is still with the decision-makers getting the advice from the AI and AI users.

    “The practice says that the algorithms and their developers are not responsible, but the responsibility is still with the decision-makers getting the advice from the AI and AI users.”

  • The paper relies on this without support.

    Oil and gas companies are not known for their agile, lean, and bottom-up development approaches but for strict jurisdiction division and waterfall processes and procedures.

    “Oil and gas companies are not known for their agile, lean, and bottom-up development approaches but for strict jurisdiction division and waterfall processes and procedures -that has to be changed.”

  • The paper relies on this without support.

    We have very few sources of geophysical and production data, and they are of questionable quality.

    “We have very few sources of geophysical and production data, and they are of questionable quality.”

  • The paper relies on this without support.

    The small data learning attracts serious attention of researchers worldwide, but there is no substantial progress as of now.

    “The small data learning attracts the serious attention of researchers worldwide, but there is no substantial progress as of now.”

  • The paper relies on this without support.

    The ability to question everything and efficiently experiment with data allows AI-born startups to attract attention and record investments.

    “The ability to question everything and efficiently experiment with data allows AI-born startups to attract attention and record investments.”

  • The paper relies on this without support.

    Oil and gas fields generate large amounts of raw data.

    “Oil and gas fields generate large amounts of raw data.”

  • The paper relies on this without support.

    AI is born in open and collaborative environment as a consequence of academia being a leading force in AI research for decades, almost without any business influences.

    “Artifical intelligence is born in open and collaborative environment as a consequence of academia being a leading force in AI research for decades, almost without any business influences.”

  • The paper relies on this without support.

    Even though many companies announce bringing some of their data to the open-source and claim the necessity of cross-company and cross-border data sharing, the reality is rather pessimistic now.

    “Even though many companies announce bringing some of their data to the open-source and claim the necessity of cross-company and cross-border data sharing, the reality is rather pessimistic now.”
16 of 16

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Cites

Flagged first, then by citations

Research works printed but not linked: 8

  1. 2018, Notes from the AI frontier: Modeling the impact of AI on the world economy, Discuss Pap
  2. R. Kohli and S. Johnson, “Digital transformation in latecomer,” vol. 10, no. 4, pp. 141–156, 2011.
  3. Kane, 2015, Strategy, not technology, drives digital transformation, MIT Sloan Manag Rev
  4. Nolen-Hoeksema, 2013, Elements of hydraulic fracturing, Oilfield Rev
  5. Ng, 2016, What artificial intelligence can and can’t do right now, Harv Bus Rev, 9
  6. Kiron, 2019, Strategy for and with AI, MIT Sloan Manag Rev, 60, 30
  7. Gazprom Neft, “The Gazprom Neft Board of Directors confirms the company's digital transformation strategy to 2030,” 2019. https://www.gazprom-neft.com/press-center/news/the-gazprom-neft-board-of-directors-confirms-the-company-s-digital-transformation-strategy-to-2030/ (Accessed 23 Apr 2020).
  8. Fountaine, 2019, Building the AI-Powered organization, Harv Bus Rev, 63

Cited by

Sorted by citations

Related

Authors
Koroteev, D., & Tekic, Z.
Affiliations
Skolkovo Institute of Science and Technology (RU)
Corresponding author
Koroteev, Dmitry
ORCID
Koroteev, Dmitry
Venue
Energy and AI
Publisher
Elsevier BV
Published
2021-03
Volume
3
Pages
100041
ISSN
2666-5468 (ISSN-L)
DOI
10.1016/j.egyai.2020.100041
OpenAlex
W3117732367
Semantic Scholar
77f65c1beb5868a8a9960491e818b3811b4325d9
Cited by
285
References
48
Citations by year
2026: 47 · 2025: 74 · 2024: 57 · 2023: 50 · 2022: 35 · 2021: 22
FWCI
10.51
Type
journal article
Language
en
Open access
Licence
Locations
  • Energy and AI (publishedVersion, cc-by-nc-nd)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
  • Skolkovo Institute of Science and Technology
  • Saudi Aramco
  • Samsung
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › Artificial Intelligence Applications
  • Social Sciences › Decision Sciences › Management Science and Operations Research › Impact of AI and Big Data on Business and Society
  • Physical Sciences › Engineering › Ocean Engineering › Oil and Gas Production Techniques
Keywords
artificial intelligence, oil and gas industry, upstream oil and gas, artificial intelligence applications, data availability, AI adoption barriers, de-risking, future scenarios, digitalization of energy

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-06.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-06.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review