OpenAI wants to build the technology that will finally create a computer that can converse in a way that is indistinguishable humans.
The real challenge is how the researchers will protect the AI from venturing over to the darker side of the freewheeling forum....OpenAI doesn’t want to become the second coming of Microsoft’s chatbot Tay, which turned racist and genocidal after interacting with people on Twitter.
OpenAI will have to have a really good classifier for what qualifies as good content and isn't just trolling, shitposting, etc. But if they can filter for that, then they will have created a really high quality dataset that can help move the ball forward with regards to conversational AI.