AIs can appear as a black box to us. What is actually happening behind the scenes? What is being done with our inputs?
Low hanging fruit would be to acknowledge a command that is given multiple times. This can be done right after the command is given or even as an update email a few days after. However, it should be personalized for the user. A good example of such a personalized email is the Google Maps timeline.
For voice, a few examples of this are:
- Bixby saying “Hey… you’ve asked to ‘Turn On Hotspot’ a few times… next time, just say ‘Hotspot On’”
- Alexa saying “You’ve asked to play You Don’t Mess with Jim 20 times last month. Next time, just say ‘Play my favorite’”
- Google Assistant saying “You’ve ask to call the same number a bunch of times. If you need to call again, just say ‘Call last number’”.
We need to know about the new things our AI assistants know in order to use they better.