Report on Current Developments in AI for Social Good and Conversational AI
General Direction of the Field
The recent developments in the field of AI for Social Good (AI4SG) and Conversational AI indicate a shift towards more inclusive, community-centered, and ethically grounded approaches. Researchers are increasingly recognizing the importance of integrating the perspectives and needs of community organizations into AI development processes. This shift is driven by the realization that the success of AI initiatives aimed at social good hinges on the active involvement and leadership of the communities they aim to serve. The emphasis on co-leadership and data co-liberation suggests a move away from top-down, technocratic approaches towards more collaborative and participatory models.
In the realm of Conversational AI, there is a growing focus on the nuances of user interactions and the limitations of third-party evaluations. The field is advancing towards more accurate and comprehensive evaluation methods that capture the subtleties of human-chatbot interactions. This includes advocating for the integration of direct user feedback into the evaluation process, which is seen as crucial for enhancing system development and user satisfaction.
Noteworthy Developments
AI for Social Good Partnerships: The paper that explores the perspectives of community organization members on their partnerships with AI teams is particularly noteworthy. It highlights the importance of co-leadership and data co-liberation in fostering more effective and ethical AI development.
Evaluation of Conversational Chatbots: The study comparing online and offline evaluation methods for conversational chatbots stands out for its insights into the limitations of third-party evaluations and the benefits of direct user interaction feedback.
