SaTC: CORE: Small: Study, Detection and Containment of Influence Campaigns

影响力行动越来越多地依赖社交网络来传播和放大具有重大社会影响的错误信息和仇恨言论。尽管社交网络实施政策并试图阻止此类不良内容和推广不良内容的账户的传播，但只有在准确理解影响力运营资源和策略以及用户信念和分发不良内容的原因的情况下，它们的努力才能取得成功。关于影响力运作的现有知识是通过缺乏科学方法的有限新闻调查和对社交网络的法医分析收集的，这越来越限制了研究人员的访问。此外，有关用户与不良内容交互的许多知识来自实验室实验，这些实验并不能准确地模拟真实的生活行为。该项目引入了一种新的方法来研究影响力活动，并提高我们对它们构成的威胁的理解，并将利用收集到的见解来开发检测和减少活动范围和影响的解决方案。该项目有可能最大限度地减少不良内容的攻击面，并影响缺乏识别和安全应对此类内容所需背景的脆弱社交网络用户的活动。开发的解决方案可能会进一步减少偏见的人谁分发不良内容的看法，并给他们一种公平感对内容moderationtechniques.This项目建立在论文的努力，研究和包含影响力的运动需要涉及个人谁分发不良内容。为了实现这一目标，项目团队设计并进行知情同意研究，参与者在野外接触不良内容，以了解他们的目标，动机，资源，策略和看法。该项目进一步开发了一个框架，以检测影响力活动，将不受欢迎的内容和发布这些内容的帐户归因于组织者，并描述其策略。该团队还设计干预措施，使选择性受众能够减少不良内容的影响和影响，分配干预结果的责任，并向对手掩盖细节。该项目进一步制定协议，以招募相关参与者，对调查结果进行三角分析，并根据参与者在野外接触和传播不良内容的情况评估解决方案。该奖项反映了NSF的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Influence operations are increasingly relying on social networks to distribute and amplify misinformation and hate speech with significant societal impact. Although social networks implement policies and attempt to stem the spread of such undesirable content and the accounts that promote it, their efforts can only succeed given an accurate understanding of influence operations resources and strategies, and of user beliefs and reasons to distribute undesirable content. Existing knowledge of influence operations has been collected through limited journalistic investigations that lack scientific method, and forensic analysis of social networks that increasingly limit researcher access. Further, much knowledge of user interaction with undesirable content comes from lab experiments that do not accurately model real life behaviors. This project introduces a novel approach to study influence campaigns and improve our understanding of the threats they pose, and will leverage gleaned insights to develop solutions that detect and reduce the reach and impact of campaigns. This project has the potential to minimize the attack surface to undesirable content and influence campaigns for vulnerable social network users who lack the background required to recognize and safely react to such content. Developed solutions may further reduce perception of bias for people who distribute undesirable content and give them a sense of fairness toward content moderation techniques.This project builds on the thesis that efforts to study and contain influence campaigns need to involve the individuals who distribute undesirable content. To achieve this aim, the project team designs and conducts informed-consent studies with participants’ in-the-wild exposure to undesirable content, to understand their goals, motivation, resources, strategies, and perceptions. The project further develops a framework to detect influence campaigns, attribute undesirable content and the accounts that post it to their organizers, and characterize their strategies. The team also designs interventions that empower selective audiences to reduce the reach and impact of undesirable content, distribute responsibility for intervention outcomes, and obscure details from adversaries. The project further develops protocols to recruit relevant participants, triangulate findings, and evaluate solutions in the context of participants’ in-the-wild exposure to and distribution of undesirable content.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.