Alexa goes down the conversational rabbit hole Meczyki.Net

On stage again: Mars This week, Amazon demonstrated a developing Alexa feature meant to mimic the flow of natural language. The interaction between two human beings rarely follows any predetermined structure. It goes to strange and unexpected places. One topic drives debate into another, as participants inject their own life experience.

In one demo, the conversation about trees turns to one about hiking and parks. In terms of the company’s AI, Alexa’s senior vice president and lead scientist, Rohit Prasad, refers to the phenomenon as “discovery of conversation.” It’s not a proper name for a proper attribute, of course. There’s no switch that would suddenly be flipped to enable conversations overnight. Rather, it is part of an evolving notion of how Alexa can interact with users in a more human – or perhaps, more human – way.

Smart assistants like Alexa have traditionally provided a much simpler question-and-response model. Ask Alexa about the weather, and Alexa tells you the weather in a predetermined area. Ask him for an A’s score (or, honestly, maybe not), and Alexa tells you an A’s score. It’s a live conversation, no different than typing a question into a search engine. But, again, real-world conversations rarely go that way.

“Alexa gets a whole slew of questions that are a lot of information. When they are questions, you can imagine they are not point questions,” Prasad told Meczyki.Net in an interaction at the event. “They’re really about what the customer wants to know more about. What is going on in our mind right now is what is happening with inflation. We get a lot of requests like this from Alexa, and it gives you that kind of exploration experience.

Such interactive features, however, are the way things do to a home assistant like Alexa. Eight years after Amazon launched, Assistant is still learning — collecting data and determining the best ways to interact with consumers. Even when something reaches the point where Amazon is ready to show it off on the main platform, change is still needed.

“Alexa needs to be an expert on many subjects,” Prasad explained. “It’s a big paradigm change, and it takes a while to get that kind of expertise. It’s going to be a journey, and with our customers’ interactions, it won’t be like Alexa will know everything. But these questions will lead to more explorations.” where you can do something you didn’t think you were.”

It was a head-turner to see the word “Sympathy” in big bold letters on stage behind Prasad – though, perhaps not as much as what came next.

There are a few straight-up scenarios where the concept of empathy may or should be equally applicable when interacting with humans and smart assistants. Take, for example, the ability to read social cues. It’s a skill that we pick up on through experience—the ability to read the sometimes subtle language of faces and bodies. Emotional intelligence for Alexa is a notion Rashad has been discussing for years. It begins with changing the tone of the assistant to respond in a way that expresses happiness or disappointment.

The second aspect is determining the emotion of a human speaker, a concept the company has been working to perfect for many years. It’s the work that manifests itself in a variety of ways, including the 2020 debut of the company’s controversial wearable Halo, which offers a feature called Tone that’s designed to “analyze the energy and positivity in a customer’s voice.” so that they can understand how they sound to others and improve their communication and relationships.”

“I think both empathy and influence are well-known ways to communicate when it comes to building relationships,” Prasad said. “Alexa may not be tone-deaf for your emotional state. If you walk in and you’re not in a happy mood, it’s hard to say what you should do. Someone who knows you well is different This is a very high bar for AI, but it is something you cannot ignore.

The executive notes that Alexa has already become a kind of companion for some users — especially among the older demographic. A more conversational approach will likely only increase that incidence. In Astro’s demos this week, the company often referred to the home robot as filling an almost pet-like task in the house. However, such assumptions have their limits.

“It should not hide the fact that it is AI,” Prasad said. “When It Comes” [where] It’s indistinguishable – from which we are so far – it should still be very transparent.”

A later video demonstrated an impressive new voice synthesis technique that uses a minute’s worth of audio to create a solid approximation of the person speaking. It features a grandmother’s voice reading “The Wizard of Oz” to her grandson. The idea of ​​remembering loved ones through machine learning is not entirely new. For example, companies like MyHeritage are using the technology to animate images of deceased relatives. But these scenarios always – and understandably – raise some hacks.

Prasad quickly pointed out that the demo was more of a proof of concept, highlighting the underlying voice technologies.

“It was more about the technology,” he explained. “We are a very customer-obsessed science company. We want our science to mean something to customers. Unlike a lot of things where generation and synthesis have been used without the right door, it seems that a customer Would love to. We have to give them the right set of controls, including who’s voice it is.”

With that in mind, there’s no timeline for such a feature — if, in fact, such a feature would actually exist on Alexa. However, the exec notes that the technology that will power it is very much up and running at Amazon Labs. However, again, if it does come through, it will require some of the above transparency.

“Unlike deepfakes, if you are transparent about what it is being used for, have a clear decision maker and the customer is in control of their data and what they want to use it for, I think it There is the right set of steps,” Prasad explained. “It wasn’t about ‘dead grandma.’ Grandma is alive in this, just to be very clear about it.”

When asked what Alexa might look like 10 to 15 years in the future, Prasad explains that it’s all about choice – albeit more than offering a flexible computing platform for users with individual and unique personalities. There is less to add.

“It should be able to do whatever you want,” he said. “It’s not just through voice; it’s intelligence at the right time, that’s where ambient intelligence comes in. It should actively help you in some cases and anticipate your need. This is where we do conversational exploration.” Whatever you look for—imagine how much time you spend booking a vacation [when you don’t] is a travel agent. Imagine how much time you spend buying the camera or TV you want. Anything that requires you to spend time searching should go by very fast.”