Re:Why UX Design for Machine Learning Matters
by virgil_disgr4ce on Monday May 22, 2017 @11:07AM
(#14394163)
As long as there are users, it will be necessary (or desirable, at the very very least) to design experiences for them. The assumption you make is that "talking to a humanlike intelligence is so natural as to be trivial." But salespeople, for instance, still need to be trained in how best to talk to their customers, what to say, what not to say, etc..
Re:Why UX Design for Machine Learning Matters
by ThomPete on Monday May 22, 2017 @11:12AM
(#14394220)
To the extent that salespeople are relevant and using ML nothing UX wise has changed. Same rules still apply.
My assumption is that many of the benefits of ML can't be used by humans but have to be used by other machines to create a final "useful to human" output.
Machine learning will mostly remove the need for UX expertise as it will allow us to worry less about manually interacting with machines as a way to get them to do things. Instead ML will do all the in between interactions behind the scenes and present us with results.
If you think about Machine Learning as a feedback loop where only the input and result become relevant then you are back to something more akin info graphics which used to be called information design.
A much bigger problem to solve UX wise is to think about how machine learning will design things and how humans will deal with the results.
https://twitter.com/Hello_World/status/861735184990961664
In the shown example the result created by ML is many times better than a human could have done but for it to work it also requires the ability to take this design and implement it into real life as the result is extremely complex.
A lot of things to explore in this space but I think we do ourselves a disservice to speak about is as UX.