So as some of you know from this blog, I've been utilising AI to help get in spoken spanish practice between lessons since, well, I don't live in a Spanish-speaking country ;). What may not be as well known is AI is becoming more common at work too and I'm finding random uses for it that I quite appreciate because it saves me a ton of time for more interesting things.
Where this collides is the topic I picked in today's practice session was "getting help from a career advisor" lol. Shockingly, once we got through some of the basics, it had some theoretically good ideas. It offered to send me some links, which I was actually almost into seeing and then... I remembered it's just the language app not the work app and as such it will NOT, in-fact, actually do anything other than correct and analyze my conversation skills. lol oops.
Also amusing to me - at the end of each feedback report it has a "cultural tip" related to the conversation. This was today's:
Where my significant protest lay is that the AI was the first to mention PMI š. I only used it back in response to the question it asked me. And, vowels are pronounced differently in Spanish so it both took me a sec to recognize what it was saying AND made me interested to know if I could say it back in a way that wouldn't get my pronunciation yelled at.
I will say, this tech is improving fast and impressively. I had a friend try it the other day for whom English is not his first language (obv I switched all the settings around) and I was really entertained listening to the English AI - we started with British by accident but I switched it to US to make it more familiar and omg was it ever a riot. Very definitely a real feeling and sounding conversation to what you'd hear on the street. It still has a compartively short memory, but far less so than it used to, so it does feel like an actual conversation now.
The only part I consistently don't like, and this one actually drives me insane. It's effectively taking dictation and then you get your feedback from that - great in that you can see what it heard (identifies pronunciation issues) etc. Less great in that I *often* get feedback that I spelt something wrong because I missed a silent letter. A SILENT letter. I was *supposed* to miss it while speaking. The AI typed it. And I know, logically, that I can just ignore that feedback and move on. But man it annoys me every single friggin time. lol logic is not welcome here! Also, potentially because it would be one simple IF statement to solve -- if the user didn't physically type the sentence, no feedback regarding silent letters that weren't typed. But - maybe it'll learn that eventually.
Living in interesting times - I do wonder sometimes what this chapter will look like in HS History texts ;)
Posts
