It’s a future we’ve long imagined but when confronted with an Alexa that doesn’t need a watchword to activate it and can hold a conversation as well as your coworker, friend, or a casual acquaintance, it’s shocking.
True conversation voice assistance is about to become a real thing. Not today, but in the very near future thanks to, naturally, generative AI and large language models (LLM). Amazon showed us on Wednesday (September 20) a new experimental Alexa that zips miles beyond its early stilted response days and easily laps the Alexa of 2023.
Granted my experience with this new, powerful Alexa was limited to what I could see and hear in both canned video and a live demo Amazon‘s outgoing SVP of Devices Dame Limp showed us.
Even so, this is not the Alexa you or I know. It speaks clearly and without that clipped halting manner we’ve come to expect and grudgingly accept. It knows the context and has a memory.
To enable the new Alexa, Dave Limp started with “Alexa, let’s chat,” and from that moment forward, Limp was having an open conversation with Alexa where he didn’t have to repeat “Alexa.”
Alexa responded, ” Okay, what can I help you with”?
Limp sort of redirected and said, “How are you?’
That’s the kind of response that might normally throw Alexa off, sending the voice assistant into a dead-end of confusion.
This time, though, the experimental Alexa immediately responded with, “I’m doing well, I’m very happy to be able to help so many people every day by chatting with them.”
The conversation continued this way for a few more moments with Limp asking about his favorite football team and how the quarterback performed in the last game. It was an unexceptional give and take, a normal conversation with an absolutely not normal conversationalist named Alexa.
Alexa’s voice sounds, well, not entirely human but closer than ever. Think the movie Her but without the sexual innuendo. That now decade-old film, released a year before Alexa, is the disturbing story of a man played by Joaquin Phoenix who falls in love with an AI voiced by Scarlett Johansson. It’s a disturbing romance that probably painted too dark a picture of a future that’s now more like our reality.
To be clear, even this experimental Alexa isn’t quite at that level, yet.
Still, because Limp didn’t have to repeat the watchword, the flow of conversation continued as if Alexa was almost human. Improved speech synthesis is set to radically alter the quality of Alexa’s voice. Ultimately, I bet it will fool many people in blind audio tests.
The humanness comes not only from the sound of Alexa’s voice but from the emotion. Alexa now sounds connected to the moment.
Alexa’s voice sounds, well, not entirely human but closer than ever. Think the movie Her but without the sexual innuendo.
It wasn’t a perfect demo. Limp had to repeat himself a few times for Alexa to respond, which is about what you’d expect from experimental technology.
Alexa’s more personality-filled voice and context awareness are only part of this upcoming transformation. Alexa will eventually recognize you when you approach an Echo Show and then pick up a conversation where you left off.
Right, I know, that sounds excellent and maybe a bit creepy. Yes, to make this work you would have to opt into visual identification so the Echo Show’s camera can see you coming and basically access the data about your last conversation.
I actually like this idea since it’s not much different than when my wife starts talking to me about something we were chatting about yesterday. Her reasonable expectation is that I remember and that I can continue the conversation. Granted, I sometimes fail in this regard, but why should a computer ever fail this way?
The scenario is you walk up to the Echo and start speaking. If you say “What time is the game?” Alexa will remember which team you like and that it’s probably related to today’s game and not one next week. The conversation could flow from there.
We don’t know much about how Amazon trained its LLM, though it did mention “Responsible AI.” Let’s just hope that’s more than lip service and that Amazon has trained the AI on diverse voices and also not with snippets of customers’ recorded utterances.
We won’t see this new Alexa until probably 2024, which not coincidentally will mark Alexa and the Echo’s 10th anniversary. What better time to release a completely new Echo and this game-changing Alexa voice assistant?
I have a feeling that this is the update that puts Siri and Google Assistant on notice. Google may be further along in this space, but Apple is miles behind in making Siri truly conversant. If the next Alexa is as good as I expect, Apple may have to submit Siri for emergency brain surgery.
Even if the first blush of this LLM-powered Alexa doesn’t fulfill today’s promise, I have no doubt it soon will. Sure, our excitement over Generative AI and LLM is mixed with generous helpings of fear about what the future may bring but this is just another signal that the AI train is fast approaching and it will never stop.
You might also like
