"One day the AIs are going to look back on us the same way we look at fossil skeletons on the plains of Africa.”
– Nathan in Ex-Machina

I have written about working with AI and machine leaning, including my experience with conversational AI. In case I am being too jargon-ish, conversational AI is a type of artificial intelligence that enables regular people to interact with computer applications the way they would with other humans. I have a productivity tool in mind which will require this ability in a machine. I periodically check in with these tools to see how far they have come and right now the interface making the most progress is online chat. It's impressive. Last week I used one to write the first section of my newsletter.
While having the robot write for me sounds impressive, the chatbot's ability to generate new content is suspect. I teach a class to high school students and when I read what the robot spits out it reminds me of them. When a student is giving a lackluster effort to an assignment it will be full of generalizations, filler language, and they will talk around the prompt. The non-answer answer. The AI chatbots do the same thing. It makes middling non-specific content. Kind of like me.
What I'm most impressed with is the robot's ability in computing my prompt's intent. This is what I need it for. I want a bot to pick up on what the user is trying to communicate then check for understanding before committing the input into use. For instance, if the user gives the bot a random list of priorities, I want the bot to do what I do: make a list, get confirmation on the list's contents, then prompt the user to prioritize the list some more.
Oh, it needs one other thing. I need to be able to program it by describing the process and output in natural language too. I mean, I'm not going to learn how to program or hire a programmer or anything.
It would help if the interface is a little like Alicia Vikander in Ex-Machina too.
Not there yet.
|