Multimodal Virtual Avatars for Investigative Interviews with Children

Abstract

In this article, we present our ongoing work in the field of training police officers who conduct interviews with abused children. The objectives in this context are to protect vulnerable children from abuse, facilitate prosecution of offenders, and ensure that innocent adults are not accused of criminal acts. There is therefore a need for more data that can be used for improved interviewer training to equip police with the skills to conduct high-quality interviews. To support this important task, we propose to research a training program that utilizes different system components and multimodal data from the field of artificial intelligence such as chatbots, generation of visual content, text-to-speech, and speech-to-text. This program will be able to generate an almost unlimited amount of interview and also training data. The goal of combining all these different technologies and datatypes is to create an immersive and interactive child avatar that responds in a realistic way, to help to support the training of police interviewers, but can also produce synthetic data of interview situations that can be used to solve different problems in the same domain.

Effectiveness and efficacy of a culturally adapted training program for child interviewers in Chile.

“…For Chile in particular, the adapted program tested in this work offers an evidence-based solution to the mandatory training for child interviewers that the new law about video-recorded interviewing of child victims establishes (Navarro et al, 2015). The current evaluation demonstrates that as with other populations (Baugerud et al, 2021;Brubacher et al, 2022;Kask et al, 2022;, when delivered in accordance with effective learning principles (i.e., spaced learning, multiple practice opportunities with personalized feedback), e-learning can be an effective strategy for foundational interviewing training in Chile. Therefore, with the addition of a module about relevant legal knowledge (e.g., the law of video-recorded interviews and the investigation of sex crimes in Chile), which could be delivered online or in a blended modality (involving online and face-to-face components; Brubacher et al, 2022;Lawrie et al, 2021), the current locally tested training program can be easily adapted to accomplish the requirements of an initial specialized training course that the new regulation in Chile defines.…”

“…Such programs incorporated spaced learning, multiple practice opportunities, and individualized feedback by expert instructors (Heidt et al, 2016; Lamb, 2016; Rischke et al, 2011). More recently, researchers have explored the utility of e-learning tools (e.g., remote mock interviews, e-simulations) to train child interviewers with promising results (e.g., Baugerud et al, 2021; Kask et al, 2022; Powell et al, 2016). Yet developments in this field are not as accessible for low- and middle-income countries where specialized training for interviewers is often limited or nonexistent (Navarro et al, 2015; Smith et al, 2019) and interviewing practices are not based on evidence (e.g., Kwakye-Nuako, 2021; Raeesi et al, 2017).…”

Synthesizing a Talking Child Avatar to Train Interviewers Working with Maltreated Children

“…After reading a number of publications [94][95][96], we turned to audio-to-video translation methods using GANs for realistic avatar synthesis which were popularized when used to generate a fake video of Obama (https://www.youtube.com/watch?v=cQ54GDm1eL0, accessed on 24 May 2022) [63,97]. Following an examination of the two approaches of ObamaNet [97] and the method proposed by Suwajanakorn et al [63] in our previous work [98], we experimented with the ICface [74], a lightweight model for face animators that is driven by human interpretable control signals. This method belongs to video-driven talking-head generation (a.k.a., face reenactment) because it uses another video as input to regulate expression, pose, mouth, eye, and eyebrows movements.…”

Enhancing questioning skills through child avatar chatbot training with feedback

“…In our ongoing work, we develop conversational child avatars, combining knowledge from developmental psychology, educational psychology, and advanced technology within AI and real-time systems. Early 2-dimensional and 3-dimensional (virtual reality) versions of child avatars that respond dynamically to trainees’ questions have been developed, and stepwise and component-by-component testing are in progress ( Baugerud et al, 2021 ; Salehi et al, 2022 ; Hassan et al, 2022a , b ). In the current study, we are testing a large generative pre-trained transformer (GPT-3), a language model ( Brown et al, 2020 ), as a child avatar chatbot to use for the training of questioning skills in an interviewing context.…”