Nextdoor is a social network for a community of neighbors where you can get local advice, buy and sell things, borrow tools, learn about local businesses and read the news. The service presented a new function that uses generative artificial intelligence. An in-app assistant will help users rephrase “potentially unfriendly” posts. The innovation should become available “over the next several weeks.”
It’s not the first time the company has experimented with ways to remind users to keep conversations “neighborly.” Nextdoor tried to combat the perception of the platform as a toxic place by sending out “kindness reminders” in 2019. And last year, the service introduced pop-ups that urge users to be more empathetic. The app also tries to combat racism and prevent heated political debates.
The new assistant, instead of warning reminders, will advise users to “rephrase a potentially unkind comments” and suggest new wording for the post. “It’s great to tell people, ‘hey, be a little bit more constructive,’ or ‘you don’t always have to respond,’” Nextdoor CEO Sarah Friar says. “But now you’re actually helping them reframe it in a way they might not have thought of.”
Messages written by the assistant are optional, users are free to make changes to the proposed text. The assistant relies on the same OpenAI models as ChatGPT. According to Friar, it was trained on “all of the data we’ve collected over our almost 10 years of life.” This allows the assistant to provide more personalized suggestions to users. The company also gave an example where the assistant suggests changing a phrase to increase the chances of getting a better response to it.
I would like to add that the Simpsons, as always, knew everything – in 2016, the episode “The Girl Code” was released, in which Lisa Simpson created an AI assistant that analyzes user comments on social networks and predicts their consequences. It is quite possible that the developers of Nextdoor were inspired by this idea.