據鉆機地帶網站6月12日報道，“生成式人工智能將對各行各業產生深遠影響”——亞馬遜網絡服務公司（AWS）的能源企業技術專家侯賽因·謝爾（Hussein Shel）表示，這正是該公司所相信的。他表示，20多年來，亞馬遜在面向客戶的服務和內部運營的人工智能和機器學習的開發和部署方面投入了大量資金。

謝爾告訴鉆機地帶網站稱，我們現在將看到機器學習的下一波廣泛應用，包括能源行業在內的每一個客戶體驗和應用都有機會通過生成人工智能進行重塑。他補充道，AWS將幫助推動下一波浪潮，讓客戶在技術堆棧的所有三層(包括基礎設施、機器學習工具和專用人工智能服務)中輕松、實用、經濟地使用生成式人工智能。

在談到生成式人工智能在能源行業的一些應用和好處時，謝爾概述說，AWS認為該技術在提高運營效率、減少健康和安全風險、增強客戶體驗、最大限度地減少與能源生產相關的排放以及加速能源轉型方面發揮著關鍵作用。

首先，生成式人工智能可以在解決運營現場安全問題方面發揮關鍵作用。

能源運營通常發生在偏遠的，有時是有害的和危險的環境中。該行業長期以來一直在尋求有助于減少前往現場的解決方案，這與減少工人的健康和安全暴露直接相關。生成式人工智能可以幫助行業朝著這一目標取得重大進展。現場攝像機的圖像可以發送到生成式人工智能應用程序，該應用程序可以掃描潛在的安全風險，例如導致氣體泄漏的故障閥門。

其次，生成式人工智能將有利于油藏建模。

他補充說，生成式人工智能模型可以通過生成可以模擬油藏行為的合成油藏模型來用于油藏建模。GAN是一種流行的生成式人工智能技術，用于生成合成油藏模型。GAN的生成器網絡經過訓練，生成與真實油藏相似的合成油藏模型，而鑒別器網絡經過訓練，可以區分真實油藏模型和合成油藏模型。

謝爾稱，一旦生成模型經過訓練，就可以生成大量的合成油藏模型，用于油藏模擬和優化，減少不確定性，提高油氣產量預測。這些儲層模型還可以用于其他對深地探索至關重要的能源應用，如地熱和碳捕獲和儲存。

最后，基于生成式人工智能的數據搜索將大大提速。

他表示，數據訪問是能源行業尋求克服的一個持續挑戰，特別是考慮到其大部分數據都是幾十年前的，并且存在各種系統和格式。

他補充說，例如，油氣公司在整個工作流程中以不同格式創建了數十年的文件，即pdf、演示文稿、報告、備忘錄、測井日志、word文檔，尋找有用的信息需要花費相當多的時間。

謝爾繼續道，據一家排名前五的運營商稱，工程師們將60%的時間用于搜索信息。通過索引增強基于生成人工智能的解決方案來獲取所有這些文檔，可以極大地改善數據訪問，從而更快地做出更好的決策。

當被問及是否認為所有的石油和天然氣公司都將在未來以某種方式使用生成式人工智能時，謝爾說他認為答案是肯定的，但他補充說，要強調的是，在定義生成式人工智能對能源行業的潛在影響時，現在還處于早期階段。

謝爾告訴鉆機地帶，在AWS，我們的目標是普及生成式人工智能的使用。

他補充道，為了做到這一點，我們為客戶和合作伙伴提供他們想要使用生成式人工智能構建方式的靈活性選擇，例如使用專用機器學習基礎設施構建自己的基礎模型；利用預先訓練的基礎模型作為基礎模型來構建其應用程序；或者使用內置生成式人工智能的服務，而不需要任何基礎建模方面的專業知識。

他繼續說道，我們還提供經濟高效的基礎設施和正確的安全控制，以幫助簡化部署。

通過機器學習應用的人工智能將是我們這一代最具變革性的技術之一，“解決一些人類最具挑戰性的問題，提高人類的表現，并最大限度地提高生產力”。

謝爾概述道，因此，負責任地使用這些技術是促進持續創新的關鍵。

AWS參加了美國石油工程師協會（SPE）國際墨西哥灣沿岸分會最近在得克薩斯州休斯敦舉行的數據科學大會活動，鉆機地帶網站總裁出席了該活動。該活動被稱為SPE-GCS數據分析研究小組的年度旗艦活動，接待了來自能源和技術部門的代表。

上個月，GlobalData在發給鉆機地帶網站的一份聲明中指出，機器學習有可能改變油氣行業。

GlobalData在聲明中表示，機器學習在油氣行業是一個快速發展的領域，有可能提高油氣行業的效率、提高產量和降低成本。

在5月份發布的一份關于油氣行業機器學習的報告中，GlobalData強調了幾個“主要的參與者”，包括bp、埃克森美孚、馬來西亞國家石油公司、沙特阿美公司、殼牌公司和道達爾能源。

本月初，數據解決方案公司Prescient的創始人兼首席執行官Andy Wang在接受鉆機地帶采訪時表示，數據科學是油氣的未來。

Andy Wang強調，數據科學包括許多數據工具，包括機器學習，他指出這將是該行業未來的重要組成部分。當被問及他是否認為越來越多的石油公司會采用數據科學和機器學習時，Andy Wang對這兩個問題都做出了積極的回應。

早在2022年11月，自稱人工智能研究和部署公司、使命是“確保人工智能造福全人類”的OpenAI就推出了ChatGPT。2022年11月30日OpenAI在其網站上發布聲明稱，ChatGPT是InstructionGPT的兄弟模型，該模型經過訓練，能夠在提示中遵循指令并提供詳細的響應。

郝芬 譯自 鉆機地帶

原文如下：

Generative AI Will Have Profound Impact Across Sectors

Generative AI will have a profound impact across industries.

That’s what Amazon Web Services (AWS) believes, according to Hussein Shel, an Energy Enterprise Technologist for the company, who said Amazon has invested heavily in the development and deployment of artificial intelligence and machine learning for more than two decades for both customer-facing services and internal operations.

“We are now going to see the next wave of widespread adoption of machine learning, with the opportunity for every customer experience and application to be reinvented with generative AI, including the energy industry,” Shel told Rigzone.

“AWS will help drive this next wave by making it easy, practical, and cost-effective for customers to use generative AI in their business across all the three layers of the technology stack, including infrastructure, machine learning tools, and purpose-built AI services,” he added.

Looking at some of the applications and benefits of generative AI in the energy industry, Shel outlined that AWS sees the technology playing a pivotal role in increasing operational efficiencies, reducing health and safety exposure, enhancing customer experience, minimizing the emissions associated with energy production, and accelerating the energy transition.

“For example, generative AI could play a pivotal role in addressing operational site safety,” Shel said.

“Energy operations often occur in remote, and sometimes hazardous and risky environments. The industry has long-sought solutions that help to reduce trips to the field, which directly correlates to reduced worker health and safety exposure,” he added.

“Generative AI can help the industry make significant strides towards this goal. Images from cameras stationed at field locations can be sent to a generative AI application that could scan for potential safety risks, such as faulty valves resulting in gas leaks,” he continued.

Shel said the application could generate recommendations for personal protective equipment and tools and equipment for remedial work, highlighting that this would help to eliminate an initial trip to the field to identify issues, minimize operational downtime, and also reduce health and safety exposure.

“Another example is reservoir modeling,” Shel noted.

“Generative AI models can be used for reservoir modeling by generating synthetic reservoir models that can simulate reservoir behavior,” he added.

“GANs are a popular generative AI technique used to generate synthetic reservoir models. The generator network of the GAN is 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,” he went on to state.

once the generative model is trained, it can be used to generate a large number of synthetic reservoir models that can be used for reservoir simulation and optimization, reducing uncertainty and improving hydrocarbon production forecasting, Shel stated.

“These reservoir models can also be used for other energy applications where subsurface understanding is critical, such as geothermal and carbon capture and storage,” Shel said.

Highlighting a third example, Shel pointed out a generative AI based digital assistant.

“Data access is a continuous challenge the energy industry is looking to overcome, especially considering much of its data is decades old and sits in various systems and formats,” he said.

“Oil and gas companies, for example, have decades of documents created throughout the subsurface workflow in different formats, i.e., PDFs, presentations, reports, memos, well logs, word documents, and finding useful information takes a considerable amount of time,” he added.

“According to one of the top five operators, engineers spend 60 percent of their time searching for information. Ingesting all of those documents on a generative AI based solution augmented by an index can dramatically improve data access, which can lead to making better decisions faster,” Shel continued.

When asked if the thought all oil and gas companies will use generative AI in some way in the future, Shel said he did, but added that it’s important to stress that it’s still early days when it comes to defining the potential impact of generative AI on the energy industry.

“At AWS, our goal is to democratize the use of generative AI,” Shel told Rigzone.

“To do this, we’re providing our customers and partners with the flexibility to choose the way they want to build with generative AI, such as building their own foundation models with purpose-built machine learning infrastructure; leveraging pre-trained foundation models as base models to build their applications; or use services with built-in generative AI without requiring any specific expertise in foundation models,” he added.

“We’re also providing cost-efficient infrastructure and the correct security controls to help simplify deployment,” he continued.

The AWS representative outlined that AI applied through machine learning will be one of the most transformational technologies of our generation, “tackling some of humanity’s most challenging problems, augmenting human performance, and maximizing productivity”.

As such, responsible use of these technologies is key to fostering continued innovation, Shel outlined.

AWS took part in the Society of Petroleum Engineers (SPE) International Gulf Coast Section’s recent Data Science Convention event in Houston, Texas, which was attended by Rigzone’s President. The event, which is described as the annual flagship event of the SPE-GCS Data Analytics Study Group, hosted representatives from the energy and technology sectors.

Last month, in a statement sent to Rigzone, GlobalData noted that machine learning has the potential to transform the oil and gas industry.  

“Machine learning is a rapidly growing field in the oil and gas industry,” GlobalData said in the statement.

“Overall, machine learning has the potential to improve efficiency, increase production, and reduce costs in the oil and gas industry,” the company added.

In a report on machine learning in oil and gas published back in May, GlobalData highlighted several “key players”, including BP, ExxonMobil, Gazprom, Petronas, Rosneft, Saudi Aramco, Shell, and TotalEnergies.

Speaking to Rigzone earlier this month, Andy Wang, the Founder and Chief Executive Officer of data solutions company Prescient, said data science is the future of oil and gas.

Wang highlighted that data sciences includes many data tools, including machine learning, which he noted will be an important part of the future of the sector. When asked if he thought more and more oil companies would adopt data science, and machine learning, Wang responded positively on both counts.

Back in November 2022, OpenAI, which describes itself as an AI research and deployment company whose mission is to ensure that artificial general intelligence benefits all of humanity, introduced ChatGPT. In a statement posted on its website on November 30 last year, OpenAI said ChatGPT is a sibling model to InstructGPT, which is trained to follow an instruction in a prompt and provide a detailed response.