Transcript 0:00 [on-hold music] With advances in AI technologies, voice assistants are going to get better at every step in their process, and we're all going to appreciate it much more. 0:13 I think in a few years' time, the ability for voice-- for a voice assistant to hear and understand and respond and, or to take action on requests will make today's voice assistants look a little bit silly. 0:26 I really believe that. [upbeat music] Welcome to The Peggy Smedley Show, the podcasting voice of IoT and digital transformation, with your host, Peggy Smedley. 0:44 Hello, listeners, and welcome to today's edition of The Peggy Smedley Show. As always, I'm your host, Peggy Smedley, and what a great day to be on the radio. 0:53 I'm loving all of these great discussions we're having so far this month on voice assistants. Some of you are all in, and some of you are thinking, "No, thanks. It's too much big brother for me," and I love it. 1:08 You know, that's what we need to be talking about. So today in segment one, I'm going to zoom out a little bit and talk about AI, the tech behind these, uh, voice assistants. 1:19 Will all these advances in artificial intelligence, I guess, translate into better voice assistants down the road? And I guess, you know, more importantly, you know, that is what we want, right? I don't know. 1:32 Maybe we do, maybe we don't. But before we get even more into that, later on in the show, I'll be interviewing Ron Babich, who's the CEO and co-founder of Easy Voice, and he's gonna chime in on the discussion. 1:45 And then I'm gonna wrap up the show talking with Chris Vandersluis, who's the president and founder of HMS Software. And they've got much more to talk about in all of this in, uh, the industries that we're in right now. 1:58 So let's get started. Grand View Research, uh, latest, uh, AI research suggests the AI market will reach almost three hundred and ninety-one billion by twenty twenty-five. 2:10 So as I always like to show you, here's some more numbers about where the market's going to, uh, grow to. 2:16 And I think it's interesting because one of the key factors driving growth in this market is advances in voice recognition, which is also going to help drive the market for voice assistants. 2:29 Now, let's take a closer look at the tech behind voice assistants. How exactly do they work? 2:36 This is actually really interesting when you kinda dive into it because USC breaks it down in a really great way, something that helped me really understand it. 2:45 And it summarizes it for you in, in a simple way, and, and let me see if I can help you understand it as well. The first step is speech-to-text. 2:55 During this step, speech-to-text software breaks the speaker's speech down, uh, and I think it's kind of an interesting way into tiny bits called, uh, phonemes. 3:07 And the way they do it, the software uses the order, a combination of the text of these phonemes, to figure out what words as the speaker is speaking, w-which would mean us. 3:20 So it must account for the fact that a lot of words sound really similar. It must also account for any noise that's going on in the background that might muff up kind of the speaker's words. 3:34 Of course, it must also account for the fact that everyone pronounces the same words a little bit differently. Uh, and it's incredible if you think about it. 3:44 Now, we all have different accents, we all say things differently, um, and there's words I know I say very differently as well. 3:51 So if you think about it, the next step is text to intent, which goes a step further and figures out what the speaker actually means. 4:02 So natural language processing is kind of like magic, and that's how I look at it, because it, it feels pretty seamless to most of us most of the time. 4:12 Uh, IBM's tech DeepQA is a great example of this, and this is what I like about it, because DeepQA can figure out what someone is really asking, and then it starts coming up with potential responses to this question and creates a thred, uh, kind of a thread, I should say, uh, for each possibility. 4:32 Uh, and there would be an example how I stumble down a word, right? 4:35 So if you think if we look at it, IBM says each thread uses hundreds of algorithm, algorithms to figure out how likely each possible answer is to relevant and how relevant, and then it creates and ranks a lot of the answers. 4:54 And that's what I found really interesting as I was researching this s-segment, because the final step in all of this, according to USC, is intent of action. 5:05 So this is the f- the step that directs the voice assistant to actually complete the request or command. You know, so buy those hundred pizzas that you might want or maybe don't want. 5:17 So as we've already discussed, the commands most people currently give voice assistants are relatively simple, like asking what the weather is or traffic is like or asking for information that can be found, uh, from a simple web search, right? 5:34 The ways we look at things. 5:35 Now, someday, these tasks will become more complex, and that's what we've been talking about all month long, and we really rely on voice assistants to set appointments for us, uh, to maybe buy items, as I was just kidding about, o-on our behalf, and to control our homes. 5:54 But what I'd like to consider today is what will it take for this to reallyGet us to where we want it to be, what we really hope that these voice assistants will do. 6:07 Which step in the process I just laid out for this and for all of us is the weak link that we have to think about? Is it the technology's ability to understand language and interpret the meaning? 6:21 Is it the ability to formulate responses or answers, or is it the ability to take action on our requests? And how will advances in AI make a difference in the realm of voice assistants? 6:37 There's still a lot of issues, I think, with voice assistants like Siri and Alexa and Google Assistant in general or Cortana being able to really riddle out, 6:49 I think, what people mean when they don't say things a c- certain way, a very particular way. 6:57 But I think one problem is, and I've had this problem in my own experience, and sometimes it's hard to figure out how to phrase a request so that it's simple and clear. 7:10 Uh, and, and it's kind of funny because I was talking to one of my writers as we were preparing this radio segment who has a Google Home, which is different because I have Alexa, and her Google Home Mini smart speaker is in her living room. 7:25 And she was telling me how frustrating it's been for her to try to remember how to phrase a particular request because it won't work unless she says it a certain way, and the problem is when she's trying to use voice commands to get the live feed from her video doorbell to-- in, and to play on her television. 7:45 So that's where she gets really frustrated, and most of the time she just simply ends up giving up. And we're laughing about it, but the reality is think about that. 7:55 Really though, this is indicative of a much bigger problem when we think about how we use voice assistants. 8:03 If, if you can't get Google Assistant to do something like that, then maybe you won't even try to ask it to buy something for you. Now, have you had a similar experience? Now I can tell you this. 8:14 My husband has said repeatedly to Alexa, "Alexa, you're stupid." Now I hate that word stupid, but he does because Alexa does not respond the way he wants. So there's another example of poor Alexa. 8:28 She, she doesn't know why she's done something not the way he wants her to, but it's happened and I've heard that many times in my house, and it's kinda like, you know. And she's probably like, "I don't understand that." 8:38 Well, of course she doesn't understand that because she doesn't know what just happened, and it's a word that they're not programming into it. So I'd love to hear your frustrating stories. 8:47 I mean, what's happening with your voice assistant? I'd love you to share those on social media with me. 8:53 Tell me about-- Let's talk about these annoying fails that indicate a larger problems in the space that are holding adoption back. 9:01 I mean, that's what we need to be talking about 'cause then maybe we can help AI move forward. 9:06 You know, with advances in AI technologies, voice assistants are going to get better at every step in their process, and we're all going to appreciate it much more. 9:17 I think in a few years' time, the ability for voice-- for a voice assistant to hear and understand and respond and, or to take action on requests will make today's voice assistants look a little bit silly. 9:31 I really believe that because I-- You know, I'm a tech junkie, you know. 9:35 I love the way we've seen it, and I love the way it's advanced, and who would have thought we would all become so excited about this the way we are? So here's my opinion for today. 9:48 Voice assistants are going to get good, I mean really good, and I think we need to think about this. 9:54 Their ability to mimic human language and, and its understanding is going to open up so many doors to using AI in ways that once required humans. I mean, that's what we need to be thinking about. 10:08 The question becomes then is there such a thing as voice assistants actually becoming just too good? I mean, that raises that really interesting question. 10:19 And I think ethically speaking, do we need to know when we're interacting with a robot and when we're not? We've talked about that before. Will it matter as long as our needs are getting met in the end? 10:31 They're great questions, right? Something we need to be thinking about. But then again, here's my tip for today. 10:38 In addition to, you know, what I talked about last week in segment one, privacy and security, we also need to remember that voice assistants are AI, and one of the key issues when we talk about AI is ethics. 10:56 We can't forget that. There are bound to be some ethical questions that will come up as this technology advances and starts being used in verticals like healthcare, hospitality, education. 11:11 We need to be thinking about that. So what ethical, I think, conundrums can you think of that have to do with voice assistants? A lot come to mind as, you know, things we're ordering and healthcare. 11:24 You know, when I think about that, there are a lot, but what are you thinking about? And that's what we need to think about. 11:29 And I, and I noticed an article came out that Frontline is doing a docu- documentary that's gonna be out tonight, and I'm really curious about that documentary 'cause I went and talked to Amazon about Alexa and they told me a, a little bit about what Alexa does, and a lot of that's gonna be in the documentary about what other people and interviews with them. 11:50 So we'll be curious to see what Amazon has to say. So there's a lot of really cool things going on right now in the AI space, so we really all should be paying attention to what's gonna happen. 12:00 So again, send me your tweet. Let's continue this discussion on social media at connectedwmag. I'd love to hear it. We're gonna take a quick break. Check out our website at connectedworld.com or thepeggySmedleyShow.com. 12:12 And remember, this is The Peggy Smedley Show, the podcasting voice of IoT and digital transformation. With great technology comes great responsibility. We'll be right back after this commercial break.
