Tech companies have long lamented the rise of misinformation on their platforms, and now some are trying to make a difference.
Eyeing Facebook and other social media platforms as their next frontier, tech companies are now developing artificial intelligence to counter misinformation. One of the companies aiming to do so is Wordflow, a startup founded by two Stanford dropouts, which is targeting a November debut.
Wordflow's AI will crawl social media for content, curate the best links from different sources into a mashup and build an understanding of the quality and intent of a given piece of content.
So while it might only filter some of the content on a platform, like Twitter, it can be customized to surface important news and updates, not just silly memes.
"It could show you different kinds of interesting things: the news that you are not following because it's trivial but true, and it could also showcase relationships," cofounder Lyle Ungar said. "The best links get highlighted so people can follow."
Ungar and his colleague, Raju Palios, worked with businesses like Caterpillar and Boeing on the first use case for Wordflow. When Wordflow started gathering data for that project last summer, they encountered hundreds of thousands of pieces of content and generated "millions of human-readable sentences," and that's only the beginning. "Once the algorithms have proven themselves, we can take it to a whole new level," he said.
All this learning does not mean everything on social media will start turning into a chatroom with 24-hour-long conversations. Ungar said Wordflow will process only links and cannot necessarily change the fact that a given post contains several fake posts from different accounts.
"But we can monitor for things that are maliciously crafted," he said.
Wordflow is currently accepting sign-ups, and starting May 1, it will open to the public. The company's platform is free to use; startups that use Wordflow to power their own products pay only for Wordflow's data and technical services.