The fifth working group focused on the importance of positive examples of AI being used for good. They brainstormed various ideas, including personalized AI assistance to defend against scams and AI attacks, using AI tools to identify negative spirals in social media interactions, and creating systems that can warn against scams while discouraging their use. They also discussed the emotions they would like to promote in a post-scarcity future and the risks associated with AI being used for negative purposes. The group suggested repurposing existing AI work to enhance safety and create systems that can intervene in conversations to prevent harmful outcomes. They mentioned the idea of creating a blog to publish a proof of concept project using public language models. They talked about the timeline for completing the project, testing with human volunteers, and the potential for funding. The group also briefly mentioned the Future of Life World-Building Contest for AI as a source of inspiration. Overall, their focus was on imagining and working towards positive scenarios for AI implementation.