Recknsense

Calculating intent is harder than AI thinks

Working out a person’s intent is hard to calculate.

In a given situation, intent is made up of different factors including: the likely outcome based on experience of yourself and others, the mood of yourself or people who are affected, the outcome you see happen for someone else, and what else may be happening around you. Currently most AI is based on the first factor i.e. by using significant amounts of data we can understand a person’s intent because it proved to be true on a large scale.   

 

Taking an example like the intention of making a purchase, it may depend on the experience you have (If I buy this it will do X), your mood (I’m feeling positive about X), the outcome for someone else (I saw someone else buy it – see a recent post about intelligence and mimicking), what else is happening (there’s an offer on for X) and probably a lot more that impacts your intention to make a purchase at that time.

For AI to be get better at working out intent, it will need to take a lot more factors into account. It will need to be tuned to listen to emotion, other people that may affect decisions, and the external activities that may be taking place.

Until then we should be aware that AI’s calculation of intent in its current form may be not be as accurate or useful as we hope.

Share this post

Share on facebook
Share on twitter
Share on linkedin
Share on pinterest
Share on email
Scroll to Top