This paper revisits the Communication and Engagement Toolkit for CO₂ Capture and Storage (CCS) projects proposed by Ashworth and colleagues in collaboration with the Global CCS Institute. The paper proposes a new method for understanding the social context where CCS will be deployed based on the toolkit. In practice, the proposed method can be used to harness social data collected on the CCS project. The outcome of this application is a development of a predictive tool for gaining insight into the future, to guide strategic decisions that may enhance deployment. Methodologically, the proposed predictive tool is an artificial intelligence (AI) tool. It uses fuzzy deep neural network to develop computational ability to reason about the social behavior. The hybridization of fuzzy logic and deep neural network algorithms make the predictive tool an explainable AI system. It means that the prediction of the algorithm is interpretable using fuzzy logical rules. The practical feasibility of the proposed system has been demonstrated using an experimental sample of 198 volunteers. Their perceptions, emotions and sentiments were tested using a standard questionnaire from the literature, on a hypothetical CCS project based on 26 predictors. The generalizability of the algorithm to predict future reactions was tested on, 84 out-of-sample respondents. In the simulation experiment, we observed an approximately 90 % performance. This performance was measured when the algorithm's predictions were compared to the self- reported reactions of the out of sample subjects. The implication of the proposed tool to enhance the predictive power of the conventional CCS Communication and Engagement tool is discussed © 2020 xx. Hosting by Elsevier B.V. All rights reserved.

本文回顾了用于CO ₂的交流和参与工具包Ashworth及其同事与全球CCS研究所合作提出的捕获和存储（CCS）项目。本文提出了一种新的方法，用于基于该工具包来理解将要部署CCS的社会环境。在实践中，提出的方法可用于利用CCS项目中收集的社会数据。该应用程序的结果是开发了一种预测工具，以获取对未来的洞察力，以指导可能增强部署的战略决策。从方法上讲，所提出的预测工具是人工智能（AI）工具。它使用模糊深度神经网络来发展推理社会行为的计算能力。模糊逻辑与深度神经网络算法的混合使预测工具成为可解释的AI系统。这意味着可以使用模糊逻辑规则来解释算法的预测。使用198名志愿者的实验样本证明了该系统的实际可行性。在来自26个预测变量的假设CCS项目上，使用文献中的标准问卷调查了他们的感知，情感和情绪。在84位样本外的受访者中测试了预测未来反应的算法的通用性。在模拟实验中，我们观察到大约90％的性能。当将算法的预测与样本外受试者的自我报告反应进行比较时，就可以测量这种性能。©2020 xx，讨论了所建议的工具以增强常规CCS通信和参与工具的预测能力的含义。