SBIR Phase I: Social Platform with Machine Learning Moderation

这个小型企业创新研究(SBIR)第一阶段项目的更广泛的影响/商业潜力将是使用一种新型的人机混合调节技术，创建一个基于肯定同意原则的无滥用和无广告的社交网络。骚扰和虐待困扰着当前的社交平台。这种虐待尤其针对妇女、有色人种和LGBTQ社区，他们最容易受到这些类型的攻击。世界各地的这些人口需要一个不受骚扰的社交网络平台来交流和分享内容，并愿意支付订阅费。此外，独立的“新媒体”内容创作者还需要一个无广告、无骚扰的平台来建立受众，与粉丝或粉丝互动，控制付费内容的访问，并为他们的工作获得报酬。由于骚扰、数据挖掘或个人隐私问题而逃离其他平台的用户可能也希望有友好和安全的空间来互动和共享内容。除了订阅费之外，这项新的审核技术还有可能作为服务提供给其他社交网络企业。这个小型企业创新研究SBIR第一阶段项目采用了一种基于同意的新方法来进行社交网络审核。该项目结合了来自多个来源的数据，包括机器学习模型，以便让用户更好地控制他们希望查看的内容，并防止骚扰和滥用。通过用人工方法补充自动化方法，并让用户控制哪些审查内容和元数据来源值得信任，这一创新将减少人类审查所需的劳动力和机器方法的准确性，同时做出相对强有力的保证，最脆弱的用户不会受到虐待或骚扰。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

The broader impact/commercial potential of this Small Business Innovation Research (SBIR) Phase I project will be to create an abuse-free and ad-free social network based on principles of affirmative consent using a novel hybrid human-machine moderation technology. Harassment and abuse plague current social platforms. This abuse is particularly directed towards women, people of color, and the LGBTQ community, who are most vulnerable to these types of attacks. These demographics around the world need a harassment-free social network platform to communicate and share content and would be willing to pay a subscription fee. In addition, independent "new media" content creators also need an advertising-free and harassment-free platform to build an audience, interact with followers or fans, control access to paid content, and receive payment for their work. Users who have fled other platforms due to harassment, data mining, or personal privacy concerns may also want friendly and safe spaces to interact and share content. In addition to subscription fees, this novel moderation technology has the potential to be offered as a service to other social network businesses.This Small Business Innovation Research SBIR Phase I project takes a novel consent-based approach to social network moderation. The project combines data from multiple sources, including machine learning models, in order to give users more control over the content they wish to see and prevent harassment and abuse. By supplementing automated approaches with human ones, and putting users in control of which sources of moderation content and metadata to trust, this innovation will reduce the labor required for human moderation and the accuracy of machine approaches, all while making relatively strong guarantees that the most vulnerable users will not be exposed to abuse or harassment.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.