In short
Tech Life Podcast Episode Summary: Talking Translation
Episode Overview
- Podcast Title: Tech Life
- Episode Title: Talking Translation
- Description: This episode explores advancements in AI technology for real-time language translation, examines the Indian government's ban on online money games, and discusses the impact of AI in the workplace globally.
- Presenter: Shiona McCallum
- Producer: Tom Quinn
Key Topics Discussed
- AI in Language Translation
- Introduction to AI Translation:
- The Pixel 10 smartphone by Google features real-time language translation during phone calls.
- Capable of translating conversations in 11 languages, including English, Hindi, Russian, and Spanish.
- Testing the Feature:
- Liv McMahon tests the Pixel 10’s translation with a call to Pablo Rodriguez in Madrid.
- The technology attempts to mimic each speaker's voice while translating in real-time.
- User Experience:
- Liv highlights the comfort level during conversation but notes occasional struggles with accuracy.
- The translation happens on-device, using Google's Tensor G5 chip for real-time processing.
- Expert Insights:
- Cien Chang, a product manager at Google, explains the technology’s inner workings and emphasizes challenges such as translation quality, latency, and user privacy.
- Perspectives from Human Translators
- Interview with Vasiliki Prestige:
- Prestige shares a balanced view on AI translation, seeing it as a productivity tool rather than a job threat.
- Emphasizes the importance of context in translation, noting that AI may not replace human interpreters in specialized situations.
- Pros and Cons of AI in Translation:
- Benefits:
- Enhances communication in casual scenarios.
- Democratizes access to information.
- Concerns:
- Lack of accountability in case of errors.
- Privacy risks related to data handling.
- Potential for technological discrimination against non-Western linguistic communities.
- India's Ban on Online Money Games
- Overview of New Regulations:
- The Indian government has outlawed all forms of online games involving monetary stakes, including poker and fantasy sports.
- Violators face severe penalties, including jail time and hefty fines.
- Government's Rationale:
- Claims the ban is aimed at preventing addiction and protecting young people from financial harm.
- Industry Response:
- The gaming industry argues that it plays a responsible role and contributes significantly to the economy.
- Companies are exploring ways to remain operational by offering games without monetary rewards but face challenges in profitability.
- AI Adoption in the Workplace
- Global Study Findings:
- Dr. Mona Murshed, CEO of Generation, discusses a study showing that 65% of their alumni are using AI tools in their professional roles.
- A majority report increased job satisfaction and productivity due to AI integration.
- Barriers to AI Utilization:
- Factors such as a lack of training and uncertainty about applications hinder more widespread adoption.
- Contrary to popular belief, fear of technology is not a significant barrier—rather, access to training and clear use cases are crucial.
- Future Employment Landscape:
- The integration of AI may lead to a reconfiguration of roles, with a potential reduction in entry-level vacancies but the creation of new job opportunities as technology evolves.
Key Takeaways
- The advancements in AI language translation hold the potential to significantly alter communication across cultures, although there are limitations and concerns regarding accuracy and privacy.
- The ban on online money games in India illustrates the complexities of regulating technology while balancing economic impact and social responsibility.
- Adoption of AI in workplaces is growing, with many workers finding it enhances their productivity, but there remains a critical need for training and transparency in its applications.
Conclusion This episode of Tech Life provides an insightful look at the intersection of technology, regulation, and social impact, shedding light on both the opportunities and challenges presented by emerging AI technologies.
Written by AI. May contain mistakes. Listen to the episode to check what was said.
Transcript
Automatic transcript. May contain errors.0:00Hello and welcome to Tech Life on the BBC World Service, the programme about technology and how it affects your life. I'm Shona McCallum. This week, will calling abroad ever be the same again? We'll be testing the AI that offers live language translation on the telephone. A government crackdown on betting bans all online games in India involving money. Our correspondent in Delhi has the latest. And more AI, this time at work. Would it make your job easier? And are you allowed to use it? I talked to an expert about how the tech is being adopted.
0:56Let's start today by talking in English about how tech can potentially help us communicate in other languages. There are various tech translation tools out there, but Google has launched a smartphone which, the company says, breaks down language barriers during phone calls. The Pixel 10 features built-in AI that can translate conversations in real time, even mimicking each person's voice, allowing callers who speak different languages to communicate more naturally. We'll find out what a real human translator makes of it shortly, but First Tech Life's Liv McMahon has joined me here in the studio.
1:33Hi Liv. Hi Jonah. Now Google supplied you with a Pixel 10 phone for you to do this test. Tell us about the live voice translation feature. So live translation for phone calls is among a slew of AI features that Google sought to cram into its new phones. It's designed to let someone with a Pixel 10 make and have calls in English as well as in 10 other languages, including Hindi, Russian and Indonesian, that can be translated in real time to another user. The feature also tries to match how a person would sound speaking in a different language while translating them. Now, I wanted to test the feature out with a real phone call myself.
2:09So I arranged to call up a friend of yours, Shona Pablo Rodriguez, who speaks Spanish and lives in Madrid. And you're about to hear a recording of that test call. It hasn't been edited, so you will hear some silent moments while I'm activating the live translation option at the beginning, followed by a notification from Google. But this is what I heard when I tried to speak to a Spanish speaker in English using Pixel 10's call translate feature.
2:40Hola.
2:44Call is translated by Google AI. In each speaker's voice, audio is not saved. Hello, Pablo. How are you? Bien. ¿Qué tal? Well, how are you? And how's your day been so far? Muy bien. What have you been up to today? Muy bien. Bueno, si has estado haciendo hoy. Very well. The day is going very well. Well, today I went to work on a normal day and then home and to rest with my family, one special, for the time being. Oh, I'm glad you got to chat to Pablo. and what we heard there is exactly what you heard on your end, Liv. What was he hearing? Well, Pablo spoke to me on his own personal mobile phone, not the new Pixel 10.
3:35And so that we could demonstrate what he was hearing from me down the phone at his end, as well as hearing what he said in Spanish, he borrowed a recorder and switched his phone to speaker mode. So I'll now play you a clip from that recording. Again, it hasn't been edited. This is exactly what Pablo heard from me on his side of the call and what he said to me in Spanish with the Pixel phone doing the translating. Has the weather been nice in Madrid today? ¿Ha mejorado el clima en Madrid hoy? ¿Cómo es? Muy bien. El clima es muy agradable. Hace un día primaveral, un tiempo muy agradable. Está muy bien.
4:12It's a very pleasant weather. It's very... I'm very jealous. Estoy muy celoso. So Liv, how did you find speaking to Pablo and letting the phone do the translating? Did it feel comfortable, like a natural conversation? It certainly let me and Pablo speak to each other in our own native languages quite well and comfortably, but at times it did struggle to balance translating and matching his voice at the same time. And as a non-Spanish speaker myself, I was quite aware that, you know, you're relying on the technology to get the translation right, which AI sometimes struggles to do. So the Pixel phone, the one that you were testing, did the translation for both phones?
4:55So the translation is done on board the Pixel 10. It happens on device and the phone can connect to a two-way translation and operate it itself. It doesn't need Wi-Fi, often just Wi-Fi to download the languages in the first place. And it will work with a regular phone connection for both outgoing and incoming calls. So how does this live phone call translation technology work? Well, I've been speaking to Cien Chang, a product manager on Google's Pixel software team. He explained to me what goes on inside the phone, and it's basically possible because of their new Tensor G5 chip. Through Tensor G5, we actually have like AI models that lives on the device that in real time samples the original voice that the speaker is saying.
5:41And then we will be taking in the original audio and then translate that into the listener's language. And we will also be taking a sample of the speaker's tone. So we will be able to translate that into the voice that our users actually sound like. So when I tried out this technology, something that I found when using it was that I could almost hear a little bit of myself being reproduced in a sort of echo. and is that part of the design of this as you say that it almost reproduces your voice while the call is happening so that it can be translated on the other side to the other user oh yes so what you're hearing is so after you're speaking like after you're finished speaking sometimes you will be able to hear a faint translated voice of your own like and this is basically like this is an intended design so you to help users understand that the person that you're speaking with is still listening to your translated language and those like help you understand and carry on the conversation a little bit more normally.
6:45Something that I think we've seen, we've seen various iterations of this technology so far, you know, often with kind of snippets of voice to text and but often it's things that kind of stop and start rather than kind of translate in real time a full kind of phone call. Have there been particular challenges around developing call translation software? Yeah, so of course there are several like very interesting challenges that the team has to face. Like translation quality is of course like we want, it's very important because we want to make sure whatever the user is saying is translated correctly and accurately to the other speaker.
7:26Latency is another very important factor. So like as you're trying our feature out, you may notice that when you're speaking a longer sentences, we don't actually wait until everything is set to start translation. So we can actually start translation as soon as there is enough context for the AI models. So this like greatly speeds up the conversation and make it more natural. And finally, privacy is a very, very important factor for us. Like we want to make sure whatever the user is saying and the user's voice actually stays private within the device. And for that, we actually really have to make sure everything works on the device and everything is protected for the user.
8:09So that was C.N. Chang, product manager on Google Pixel's software team, and he was speaking from Seattle in the US. Thanks to him and also thanks to you, Liv, for testing the phone for us. Thanks, Jonah. Happy to help. Well, what does a human translator think? I caught up with Vasiliki Prestige, a London-based translator. She's also a university lecturer teaching languages. She's a council member of the Chartered Institute of Linguists and she sits on the AI committee. I asked her if she viewed AI live language translation technology with suspicion or curiosity. I wouldn't say suspicion or curiosity.
8:48I think we have to have both to be more balanced and realistic in our approach as professionals. There is no suspicion around AI in terms of whether it replaces us or not, because it, in fact, is a useful tool for translators because it is becoming something that we integrate in our processes, It's something that makes us very productive. So it has to be approached in both ways. You touched on it there. I mean, looking at live language translation technology specifically, do you worry that it could be a job killer? I mean, in the context of this new development with a Google Pixel phone, I don't think so because we need to really think about the context.
9:39let's say the scenario is I want to speak to my grandmother in Spain and I don't speak Spanish, so the phone does this for me. That's great because it enhances the communication between us. Now, myself as a professional interpreter, nobody would ever call me and give me a job and pay me to enhance communication between a grandchild and a grandmother. So that's not the job of a professional interpreter. I'm being called to interpret in a courtroom or at a high commission, an embassy, a business meeting or at a law professionals meeting. So we need to be very mindful of context. Yes, absolutely.
10:23And I'm sure that you can see both pros and cons when it comes to AI and translation. I mean, what do you see as the benefits of this technology? exactly you you're spot on there are benefits and there's advantages and disadvantages the the the benefits of it is enhancing communication where a human is not there at the right time let's say or at the right place and when you need to communicate during a very casual day-to-day scenario. Also, it is making information more available readily to people and it democratizes information. That is a huge benefit and a great use for it. On the other side, of course, there are the negatives.
11:17Of course, of course, there are a lot of negatives here because, listen, there is no accountability when we engage with a technological solution such as AI translation. If something goes wrong, if there is a severe misinterpretation, who is there to be blamed? That's one very serious issue that individuals need to have in mind. The other great aspect then we need to pay attention to is the security and privacy of data. Now, there can be several claims that the data is very safe, it doesn't go anywhere, but I really don't believe that to be true. Data has to be hosted somewhere, it is shared, it is recycled.
12:06So we need to be asking for transparency around how do these tech giants manage our data? And then there is the element of technological discrimination, because we really need to bear in mind that these technologies are developed primarily for the purposes of the Western world. So we're giving solutions to a certain group of the planet. And so what happens with the rest of the world where we don't have as much linguistic data to build technologies on? That's Vasiliki Prestige, a translator who keeps an eye on AI development, speaking to me from Athens in Greece.
13:03This is Tech Life on the BBC World Service with me, Shona McCallum. Now, if you were listening to last week's show, which I hope you were, you'll have heard Karen Howe, the author of the book called Empire of AI. Now, she talked about big tech firms wanting to open data centres in South America in places where there's a lack of fresh water. Karen also said that these facilities need the water to cool down. Well, hello to listener Peter McNabb, who's emailed us from Western Australia to say, it would be interesting to hear whether small modular nuclear reactors are viable, supplying electricity to power desalination plants.
13:41Well, it's a good question, Peter. Mini nukes, as I think they're sometimes called, are definitely an interesting solution. Let me also say hi to Jeremy, who listened to Tech Life in Uganda. Jeremy WhatsAppped us to say that he's a big fan of his Kindle e-reader. Now, Jeremy says, I love to carry all that knowledge around so easily and to stay entertained and informed. Yes, I think Kindles are great too, but I just love my paper books. There's something about leafing through pages that I really enjoy. Well, thanks to Jeremy and Peter for getting in touch. And if you'd like to contact us, we'd love to hear from you, especially if you're a woman.
14:19What do you think about the show? What tech do you use? Maybe you've got a great short story to tell about your mini adventures with tech. You can email us techlife at bbc.co.uk is our email address or send us a voice note or message on WhatsApp. The number is plus 44 330 1230 320. Include your name and where you are in the world.
14:51Some gamers in India are getting used to a new reality this week. The government has outlawed all online games that have anything to do with money. This affects traditional betting games like roulette or poker, but also online games like fantasy sports that have become popular in recent years. Before the ban, India had around 400 startups involved in real money games that generated nearly US$2.3 billion in annual taxes and supported more than 200 ,000 jobs. To find out more, I got in touch with our BBC correspondent in India, who's been following the story. Hi Tech Life, I'm Davina Gupta. I'm a correspondent in India's capital city, Delhi.
15:32Great to have you on the programme, Davina. First of all, can you outline to me what the new online money gaming laws are and what they cover and what are the punishments for breaking those laws? So we are talking about real money games. And these are games where players use online platforms to bet cash on, say, poker, fantasy sports and other games, and they're able to earn from it. Now, the law states that enabling such services is punishable now with a jail term of up to three years and a fine of at least$100 ,000. And this is the first federal law to ban online betting platforms. There have been instances where different states have taken different stands in enforcing their own bans in such games.
16:19But this is the first time that the government has come with a nationwide ban. And are we talking about games of chance which rely on luck or games of skill? Well, that's the controversial bit because the law doesn't really distinguish between the two. For example, let's talk about fantasy cricket. Now, it's hugely popular. In this game, the users actually build a virtual team based on real cricketers. And as they score points in their real matches and how they perform in those matches, these users then can get points in their games. And similarly, they can get a cash out if they score higher points.
17:00Now, there's strategy involved, if you think about it, where you got to know your players and their strengths while you're building that team. But on the day of the game, there's still a lot of it up on chance, isn't it? And courts in India have previously ruled that fantasy sports are games of skill. The Supreme Court even backed that up in 2022, which is India's top court. But now the government has said that it doesn't matter whether you think it's a game based on skill or game based on chance. They say that all such games involving money should be banned. And tell me why the Indian government have brought this in now.
17:37Well, the official line is addiction. India's IT minister has said that as many as 450 million Indians are being harmed by online betting. Most of them are young people. And he's talked about how they're losing money, falling into debt. And there have been even cases of depression and people trying to end their lives. And this is what the government claims. There hasn't been a comprehensive study in this regard. But the government is saying that it's trying to do this to save people, save lives. The gaming industry, of course, has hit back. And tell me what the gaming industry has said then, because obviously that's been big business in India for a while now.
18:23Yes, the gaming industry has said that they are responsible platforms. They have said that they contribute a lot into taxes from the revenues they earn. They provide at least 200 ,000 jobs for the country, and that's a crucial part of it. And they've said that they are not addictive. They give users chance and they give them a platform to play. So essentially, they cannot be bundled together. And they're trying to fight this back by convincing and lobbying that their real money games should be allowed. And what effect do you think it will have on online gaming companies? Well, for some of them, they are trying to keep users engaged by offering the same games without real money involved.
19:12So fantasy sports that we were talking about earlier, you could still play it, but there won't be any cash out. But the big question then is, can they survive without making money? Because many analysts are saying there will be steep job losses. So these gaming companies, and many of them are big ones earning million dollars worth of revenues. They are scrambling. For now, they have closed all of these real money games. And a lot of them, like Dream 11, for example, which has been a sponsor of cricket games in India, it has pulled out of sponsorship of similar cricket games saying that we can't afford it anymore.
19:50So they are in a scrambling mode and a survival mode, trying to find out how they can keep in this market and keep up with this change in law. And presumably they'll do everything they can to fight it then? Yes, let's talk about Head Digital Works, which is the parent company of A23, which offers online rummy and poker games. They've decided to challenge this bill. They filed a complaint in a local court in southern Indian state of Karnataka. But the centre has said that the court ruling should not come in between a legislative order. So what the government is saying is that this particular bill has been passed by both houses of parliament.
20:32So it has gone through that process of being debated and consulted and then it's been passed. And now it awaits signature of India's president, after which it becomes a law. And that cannot be changed on the court's will. So it's going to be difficult for those companies which are trying to fight this ban. That's the BBC's Davina Gupta speaking to me from Delhi.
21:03How artificial intelligence will affect jobs in the future is still a global debate. But one thing is certain, the tech will reshape the workplace in some capacity. So how are people adopting it? We've been looking at a recent study by one of the world's largest employment programmes called Generation, who have alumni across the world, from Ghana, India and Kenya to Mexico and Thailand. Here's Dr Mona Murshed, Generation's CEO, to explain more. We surveyed about 5 ,500 alumni who graduated in the last two years across our programs. And they're across 17 countries and 40 professions. And what did you discover?
21:45First, we found that 65 % of our alumni are using AI in the workplace. They are not only using it, but they are power users. About 80 % are using it every week. over a third are using it on a daily basis. In addition, what we found is about half of them are self-taught. They are essentially going out and finding their own tutorials that are online in order to be able to use these tools. The great part is that we found that the majority, you know, over 60 % are finding that using AI tools makes their job more enjoyable. They feel they're more productive, they feel that they're able to do more, they feel that they're able to focus their activities on things that aren't the quote unquote grunt work.
22:35And why do you think it is workers that don't have degrees that are racing ahead with AI adoption? In many roles today, unless you are learning these AI skills, it's very challenging to be able to progress and grow in your role. And our alumni are seeing that. I also mentioned in the report, those who were not using AI, there were some obstacles, lack of training, uncertainty about the applications, maybe not sure how to use it, but not the fear of the technology, which was interesting. Once the use case becomes apparent, you find much more uptake of AI. You know, I'll give you an example. So for those of us who have to read through large documents, it takes time to do that.
23:26But if you can get a summary or a synthesis or a comparison between three or four different texts, all of a sudden you realize, okay, wow, this can really save me time. I can figure out where I want to deep dive and so on. What we found is no one was against usage of AI, but it's just give me access to the right training, help me to understand the use case. And so we expect that even the third that weren't using AI today, that that will quickly change. So perhaps people need a reason to use AI. And clearly, as the tech advances, we start being able to do more and more. And in some sectors, even replace people in jobs completely.
24:02Do you see it creating any jobs, though? We are seeing a 20 to 30 % reduction of entry level vacancies in tech and in customer service in some geographies. Now, is that only due to AI? Is it also due to economic uncertainty? Very hard to say, but it is undoubtedly the fact that AI is playing a role. Now, will it also create new job opportunities? If it follows the path of other technologies, it will. I would though say we're right now in what I would call the messy middle, which is that employers are still figuring out what kind of employee configuration they need in order to make the most of AI tools.
24:52And so we're all still figuring it out, but it is likely that there will be some roles that AI gives birth to. We've seen it a bit with the prompt engineering, etc. But those are some examples. We'll get to, I think, much more germane ones. but there has to be a lot of focus by all of us on entry-level roles because these are so important if it takes a long time to access that role there's ample research evidence to show that that results in income scarring because it delays your ability to grow and be promoted so this is a really important moment i think in workforce history as we figure out how to make the most of this technology, but enable opportunity for everyone.
25:39That was Dr. Mona Moushed, Generations CEO.
25:49Well, it's time for me to clock off for this edition of Tech Life. Remember, if you're a woman and you listen to the programme, this is your chance to get in touch and have your say. It could be about the show or about the tech you use, or maybe you've had a mini adventure with some technology. Our email address is techlife at bbc.co.uk or you can WhatsApp us on plus 44 330 1230 320. Today's Tech Life was produced by Tom Quinn and presented by me, Shona McCallum.
From the publisher
We test the AI tech that offers live language translation on the telephone. India bans online money games. And AI in the workplace - a global study looks at who uses it, and why.
Presenter: Shiona McCallum Producer: Tom Quinn
(Image: A photo of a young woman speaking into a smartphone. Credit: pocketlight/Getty Images)




