In a latest version of the Supportive Weekly publication I wrote about why we is perhaps higher off treating AI chatbots as a software for self-service, as an alternative of as a full substitute for human assist (which they only can’t utterly obtain).
One reader, Ita, requested how we may virtually obtain that reframing. Within the newest episode of Supportive Shorts (the mini-episodes of The Supportive Podcast), I shared my reply:
The 2 components are telling folks what the instruments are for, and utilizing technical guardrails on the AI to align it with the identical functions. I share some particular examples within the podcast episode.
In case you possibly can’t get sufficient of Sixties Australian tv, right here’s the complete section:Â
‘Immediate cash’: ATM involves Australia (1969) . You may solely take out precisely $25, and your card needed to be mailed again to you after every transaction. Helpful.
ATMs didn’t completely substitute financial institution tellers then, nor do they now. We needed to learn the way they had been most helpfully used (or abused).
Now we’re all studying how AI instruments work, the capabilities and limits of huge language fashions, and the way they are often greatest managed to assist ship the shape and high quality of service we would like.
Should you’re in search of some LLM studying, listed below are two good choices:
