Polite.ai is an open source project which aims to provide a set of tools that detect negative communication exchanges to improve the way that we interact online. It isn’t a communication tool in it’s own right but it will provide a dataset and interface based on deep learning to classify real time chat.
Communication that should bring out the best invariably brings out the worst
Whilst communication has never been easier using online tools, we no longer interact and build relationships in the same way. Lacking the rich verbal and non-verbal tapestry which marks real life human conversations, the structures in our brains that are tuned to discourse and building relationships work differently. At the same time the physical incentives to avoid conflict are not present.
Trolling behaviour has developed where contributors gain perverse feedback by triggering others. This is unique to the online medium.
Negative online comments have a corrosive impact on the viability of online communities. Positive contributors are discouraged and denied a safe space to discuss big issues. We retreat into safe bubbles of people who think like us which never solves anything.
So, how can we improve things
We are still in the infancy of global online discourse. New communication tools and networks are evolving all of the time and as these start to take into account the neuroscience of communication the interfaces we use will become better tuned at helping us be the best people we can online.
Part of this landscape may well be deep learning tools to detect patterns of negative communication exchanges and nudge us in the right direction.
We don’t really have an organisation, this is just an idea that is being developed by a group of individual contributors, but you can contact Rob if you want to engage with the project…