Child Privacy Protection Engine for Smart Anthropomorphic Toys

多年来，儿童玩具变得越来越复杂，越来越多地从简单的物理产品转向通过使用软件和硬件参与数字世界的玩具。在这项研究中，智能拟人玩具被定义为一种由人形形式的物理玩具组件组成的设备，该设备通过网络和传感技术连接到具有在线服务的计算系统，以增强传统玩具（如美泰的Hello Barbie和Cognitoys Dino）的功能。这种新的模式在玩具计算中引入了面向服务的体系结构方法，在玩具（服务提供者）和儿童（服务请求者）之间。参考美国联邦贸易委员会《儿童在线隐私保护法》和欧盟《数据保护指令》的方向，本研究采用儿童为13岁以下的个体的定义。儿童用户的另一个问题是儿童的使用模式与成人的使用模式不同。在这项研究中，假设是儿童往往很少理解或尊重他们的信息的隐私，更有可能以自发的方式行事。儿童的使用行为表明，他们更愿意透露私人信息，这使得敏感数据共享问题受到极大关注。许多研究发现，拟人化玩具是有目的的，因为孩子们信任这样的设计，并感到轻松地透露私人信息。在这种环境下，儿童的在线隐私一直是一个非常令人担忧的问题，特别是当涉及儿童的私人信息并可能与其他各方共享时。例如，一项名为“谷歌玩具”的新发明引起了媒体的许多批评，因为人们对谷歌可能侵犯隐私表示担忧，特别是对他们家中的孩子。隐私可以导致儿童用户的人身安全，例如，儿童掠食者。所有这些风险都随着儿童捕食者了解儿童位置或历史位置模式的可能性而增加。虽然父母努力确保他们孩子的身体和在线安全和隐私，但在这个范例中，没有标准化的儿童保护引擎用于父母控制。父母控制是智能玩具中的一个功能，用于父母限制儿童可以提供给玩具的内容。本研究的主要目标是开发一个智能玩具的隐私感知上下文数据模型，以支持一个标准化的儿童保护框架与家长控制。该框架包括一个警报机制，通过应用文本挖掘技术来识别儿童和玩具之间的可疑对话。这项研究的潜在影响是提供一种安全的智能玩具计算交付模式，以保护加拿大的儿童。该研究团队包括2名博士。信息技术专业硕士研究生2名。

Children's toys have become increasingly sophisticated over the years, with a growing shift from simple physical products to toys that engage the digital world by the use of software and hardware. In this research, a smart anthropomorphic toy is defined as a device consisting of a physical toy component in a humanoid form that connects to a computing system with online services through networking and sensory technologies to enhance the functionality of a traditional toy such as Mattel's Hello Barbie and Cognitoys Dino. This new paradigm introduces a Service Oriented Architecture approach in toy computing between the toys (service providers) and children (service requestors). Referring to the direction of the United States Federal Trade Commission Children's Online Privacy Protection Act and the European Union Data Protection Directive, this research adopts the definition of a child as an individual under the age of 13 years old. Another concern with child users is that the usage patterns of children differ from that of an adult. In this research, the assumption is that children often have little understanding or regard for the privacy of their information, and are more likely to act in spontaneous ways. The usage behaviour of children indicates that they are more open to giving out private information, which makes issues of sensitive data sharing of great concern. Many studies found that anthropomorphic toys serve a purpose, as children trusted such designs and felt at ease disclosing private information. Online privacy for children has been a great concern in this environment, particularly when the child's private information is involved and can be potentially shared with other parties. For example, a new invention called “Google Toy” has caused many criticisms from the media as people express concern about possible privacy breaching by Google, especially to their children at home. Privacy can result in physical safety of child user, e.g., child predators. All of these risks are increased with the possibility of child predators becoming aware of the child's location or historical location patterns. While parents strive to ensure their child's physical and online safety and privacy, there is no standardized child protection engine for parental control in this paradigm. Parental control is a feature in a smart toy for the parents to restrict the content the children can provide to the toy. The main objective of this research is on developing a privacy-aware context data model for smart toys to support a standardized child protection framework with parental controls. The framework includes an alert mechanism by applying text mining techniques to identify suspicious dialogue between children and the toys. The potential impact of this research is to provide a safe smart toy computing delivery model to protect children in Canada. This research team includes 2 Ph.D. and 2 Master students in Informatoin Technology.