There are a lot of interesting applications of AI. One of them is having a conversational chatbot. OpenAI has opened its chatbot at http://chat.openai.com/chat for public review and I had some fun this morning probing it about neural networks first and then our education system.
The depth of the knowledge base is really surprising. When asked for what service might be running on an open port of 5432 on a server, it correctly identified that as PostgreSQL, a popular RDMBS. But when I asked for an exploit for port 5432, the system correctly refused it.
Obviously, there are some levels of ethics built into the chat system to not give out harmful information and guide its users in the right way. This is a pretty impressive achievement for a beta version.
OK — so, what is the point of all this probing the tool? Obviously, the tool still has some glitches (duh, it’s in beta version) and hangs some conversations. But, I feel, it can be a great support for first-level treatment of disorders, especially of a psychological nature.
If one can do these even as a business, this chatbot can make significant strides in providing much-required support to people struggling with psychological problems:
— Make a mobile app for people to use with this chatbot as the backend
— Support local language chat
— Make a directory of support professionals to help those in need
I sincerely hope AI can be put to good use in society than just for military operations.