How Oil and Gas Majors Are Leading the Way in Artificial Intelligence
There is no small irony in the fact that the fossil fuel industry, attacked today as fossils of a bygone era, is one of the biggest industries leading the charge on integrating artificial intelligence in the workplace. Besides the datedness of oil and gas as fuel sources, oil companies often have old-fashioned leadership and corporate structures. Even so, necessity is the mother of invention. With the target being drawn on the back of the oil industry growing ever greater due to rising costs and dire environmental concerns, the necessity for them to innovate has never been greater. With this in mind, it bears reviewing the way in which this transformation is already occurring within the oil and gas industries, and what that means for them as well as for the future of AI.
Different Forms of AI
The first important distinction to make is that there is not a single way, but rather a myriad of ways, in which AI is being integrated into the oil and gas industries. With the emergence of applications and the theoretical principles behind the “Internet of Things” coming ever closer to becoming an everyday working reality, artificial intelligence is setting itself up to become an ever bigger part of the workplace. Two of the most common AI innovations among oil and gas leaders at present are “intelligent robots” and “virtual assistants” (Sennaar, 2019). The former are robots that have been programmed with greater AI abilities than previous generations to help them detect hydrocarbons and improve production and efficiency while reducing the risk to workers. The latter are online-based chat platforms that help field questions from customers, allowing them to better and more quickly navigate menus and databases. ExxonMobil, Royal Dutch Shell, and Sinopec are three of the biggest oil and gas companies leading the way in AI innovation, and each of them are doing so in their own way.
AI and ExxonMobil
Around 80 universities, including MIT, have worked with ExxonMobil to help develop its current crop of AI robots that can explore the ocean floor. ExxonMobil is interested in exploring the ocean for further oil reserves – a task that could be time-consuming or dangerous for humans to perform, placing the onus on developing machines that are smart enough to explore the seafloor. Giving a boost to those efforts are former NASA designers such as Brian Williams, who helped create the AI software that helped lead to the creation of the Mars Curiosity Rover. The parallel between undersea and space exploration is not a new one, and companies such as Exxon are hoping this can translate into more effective seafloor detection capability (Techet, 2009). The impetus for that move lies in the fact that, with existent oil reserves already starting to show signs of strain in some places, there is a strong need to search for oil elsewhere. Natural seeps may fill that void. Natural seeps are the term for oil escaping from rocks on the ocean floor, and undersea oil remains one of the biggest outstanding reserves still untapped, with roughly 60 percent of the oil beneath the ocean being natural seeps.
In these natural seeps, petroleum and methane gas flow upward from sea conduits toward the surface, or else are caught up by sea currents. In the former case, they make their way to the waves at the top, forming oil slicks and releasing methane into the atmosphere. In the latter case, they settle into fallout plumes on the seafloor. The hope for companies like ExxonMobil is that, by knowing where those surface or seafloor-level oil slicks are, they’ll be able to collect them without having to drill, which can be costly and cause major environmental damage. It is in hopes of avoiding the financial, environmental, and public relations repercussions of drilling that ExxonMobil has developed this latest generation of AI-powered robots. ExxonMobil hopes these robots can be used to navigate the depths of the seafloor, identify natural seeps, and eventually collect them without the need for human involvement like drilling. AI-powered initiatives such as these have been a major focus of ExxonMobil for the past two decades, during which time it has pumped around seven billion dollars per year into research and development efforts. While the amount of money spent on initiatives such as AI robots has not been disclosed, their efforts nevertheless point to one way the oil industry is responding to both an increased scarcity of oil and increased attacks on its environmentally unsound practices.
The Future at Royal Dutch Shell
Given the volume, diverse nature, and various hours at which customers press queries, automating customer service lines has long been a focus of companies both big and small. One of the biggest obstacles to that has been the difficulty of training bots to respond intelligently and intuitively to customer queries. It’s one thing to train a bot how to respond to a branching set of questions, but quite another to do so in a way that can help with complex questions. For that reason, what’s known as the “human factor” remains a critical part of effective customer service. Yet for companies such as Royal Dutch Shell, improved AI in virtual assistants as part of chat applications are proving their effectiveness as well. In August 2015, Shell announced that they would be the first oil and gas company to implement an AI virtual assistant program. “Emma” and “Ethan” – the two AI avatars used by the companies – greet customers with natural language functions that help them answer questions and direct customers to oil and gas-related products that suit their particular needs.
The information these programs are able to provide includes the availability of a given product, the pack sizes in which certain products are offered, and general answers for questions regarding specific products, as well as general technical queries and issues. The company has hopes that the Shell Virtual Assistant will be able to expand upon the results it has already returned. According to Shell, the system can give information on 100,000 data sheets, 3,000 products, and 18,000 different pack sizes, as well as describe 16,500 characteristics of various types of lubricants, and identify 10,000 competitive products. While this Virtual Assistant system is only available for U.S. and U.K. customers so far, there are four other systems, including Shell LubeMatch, which are able to expand that AI customer service. These systems are able to communicate in 21 languages and have fielded queries from customers in over 130 countries to date.
That being said, there is still reason to be skeptical of Shell’s new chatbot features. For one thing, the technology remains new, and so its efficacy remains to be seen. This is especially true given the aforementioned difficulty of matching “the human factor” in terms of fielding customer queries. While chatbots may be able to help field some of the initial incoming volume of queries, they are not yet sufficiently developed enough to go that deep into the sales process. That means that humans still have to be involved for most of the work, which leaves the ultimate business value of these bots in question. Nevertheless, it is certainly true that applications such as these represent a step forward in the manner in which websites field customer queries and make data available to customers.
Virtual assistants expressly for the purpose of customer service are not the only AI-related area in which Shell is innovating, however. The company is also looking to automate more of its facilities, and the reasons for this are clear from both a business as well as technological perspective. In the former case, automation is the wave of the future for many industries in which manual labor plays a significant role. While this has led to serious concerns about the economic status of workers whose jobs are automated, automation stands to increase both efficiency as well as productivity. From a technical perspective, automation has the potential to remove workers from dangerous or menial tasks. This can then free them up for employment opportunities that make greater use of the human factor. Among the tasks that Shell is considering automating with AI include those that involve routine observation, data gathering, and invoicing. Its existing Virtual Assistant programs may, thus, be seen as a step in that direction.
Sinopec‘s Way Forward
Other companies are also looking at the possibilities posed by on-site automation. Sinopec has announced plans to create ten “intelligent plants” which are powered in part by AI. These plants are aimed at reducing operation costs by as much as 20 percent. Sinopec is hardly alone in this move among Chinese corporations, with Huawei announcing plans in April 2017 to collaborate with Sinopec to develop smarter manufacturing platforms. Among the capabilities at the core of these new AI systems are means by which to centralize and disseminate data. This is highly important for making on-site tasks proceed more efficiently. One major cause of slowdowns in the workplace – be it in the oil and gas industry or elsewhere – is the bureaucratic red tape and storage difficulties that accompany trying to find the right data. This can be especially costly and time-consuming when it slows down work on factory floors and work counters. Just as those other AI assistants are able to match customers with data on thousands of products, so too do companies such as Huawei and Sinopec hope that smart platforms such as these will make it easier to manage on-site factory work. In addition, these types of AI could be used to set up models by which certain sets of data could be interpreted. In doing so, companies could gain valuable insight from the analytics gleaned from these data sets, allowing them to implement even more effective solutions. There is not yet a concrete timeline for these data systems’ implementation in current or future sites, though these developments are promising.
Innovations at Gazprom
Sinopec and Huawei aren’t the only pairing of energy and tech companies looking to create more efficient, effective, and ultimately profitable energy solutions by embracing AI. Russian energy giant Gazprom announced plans in June of 2017 to collaborate with the tech titan Yandex. As part of this partnership, the companies plan to work together to make use of AI and machine learning for a variety of tasks – including drilling and completing wells, modernizing and modeling new oil refining strategies, and further optimizing Gazprom’s inner workings with other technological processes. Part of the cooperation agreement allows for a great deal of R&D flexibility and freedom. This has the potential to be good news for creative minds involved in the partnership looking to explore new options in a bid to modernize Gazprom. Among the shared points of interest between the two companies include data sharing, employee training, and technical support.
The oil and gas industry stands at a crossroads. After decades of technological stagnation and a refusal to innovate, companies are now seeing a backlash to both the antiquatedness of those policies, as well as the environmental impacts of their current methods. It is clear that, to survive, things need to change within the industry. The nature of AI in these fields is, thus, directed at delivering the kind of change that these oil and gas giants hope can affect change without entirely upsetting their current foundations. With AI-powered robots that can scour the ocean floor, companies like ExxonMobil hope to be able to make the oil scouting process more efficient, while simultaneously answering some of their critics about their environmentally-damaging practices. With so much of the earth’s untapped oil reserves still buried deep beneath the waves in rocks which give rise to natural seeps, being able to collect that oil without drilling would be an enormous positive. At the same time, automating workplaces with smarter applications that tie into the Internet of Things can help workers gain access to databases far faster than they are presently able to do, while AI bots can help improve the rate at which companies can field basic queries. The journey toward greater AI integration may be a long one, but some of the industry’s largest players are decisively moving forward with AI programs that will lay the groundwork for their continued future.
Sources
Sennaar, K. (2019). “Artificial Intelligence in Oil and Gas – Comparing the Applications of 5 Oil Giants.” Emerj.
Techet, A. (2009). “Exploring Sea, Space, & Earth: Fundamentals of Engineering Design.” MIT.
