Kerala Technology
India is a sound market for AI bots

AI chatbots are now able to generate human-like conversations, opening up new possibilities. Image: Igor Omilaev/Unsplash

India is a sound market for AI bots

Hari Kumar By Hari Kumar, on October 08, 2024
Hari Kumar By Hari Kumar, on October 08, 2024

Link to podcast generated by Google NotebookLM

In the last newsletter, we included a podcast generated by AI, and the audio impressed many of our readers with its human-like quality. The ability of AI tools like NotebookLM and ChatGPT to create such audio has led to many AI-generated conversations appearing on the internet, and some of them are real killers.

As more users start testing the audio capabilities of AI platforms like Google's NotebookLM and OpenAI's ChatGPT, it is becoming clear that AI is on the cusp of another transformation, centred around the rise of audio chatbots.

With the emergence of AI-generated audio, the way we interact with machines could become more intuitive, conversational, and even more human-like. The AI surge that began with text-based chatbots and virtual assistants has rapidly evolved into a world where AI can speak, listen, and comprehend at levels approaching human communication.

Feedback from some of our readers was also illuminating, as it revealed different ways in which users utilised this new technology. The fact that some models were fluent enough in Indian languages added to its appeal.

Speaking to the AI chatbot in Malayalam or Hindi and using it to improve their English proficiency is a popular approach among students. Not just English – one student mentioned that he found it greatly helpful in his German language education.

“It helps by explaining the grammar and suggests better words for spoken usage. It’s like having a personal friend trying to help me out,” says Mishal.

The conversational nature of it makes it easy to listen to and understand even complex subjects. There are already social media posts showing kids in Kerala saying that they upload their study material onto AI platforms and ask it to explain the key points.

The ability of AI platforms to read and analyse documents and make human-like interactions based on that takes their utility to another level.

It is not just students who are finding this function of AI useful. Robin Alex Panicker, an investor and software engineer who runs his own software firm, says he uses Google's NotebookLM to sift through the CVs he receives.

“I upload them to NotebookLM and ask it to provide me with a list of those with the specific skills we are looking for. It saves me a lot of time,” he says.

Robin, who is sceptical of the ongoing hype about AI displacing humans, says this is the kind of changes that AI will bring – essentially saving time by handling mundane and repetitive tasks.

The way New York Times podcaster Casey Newton used it proves that point as he says he uploaded the manual of his washing machine to understand the operations quickly.

Those with busy lifestyles say they use AI platforms to analyse content and provide the information in capsule form, while some use it as a personalised news podcast. They simply scan the web, select the links they want to read later, and bookmark them to the AI tool. They then listen to the audio when they have time.

The ability of these tools to generate audio that sounds human-like has impressed even those in the thick of things. In a recent video, Dan Shipper, CEO of Every, expressed excitement about the capabilities of OpenAI’s new audio tool and its ability to help users get through books on difficult subjects.

“Computers can talk to us now. And if we use them correctly, they can help us learn more about our world and about ourselves, in a way that feels as natural as a conversation with a friend,” Shipper wrote in his newsletter.

Andrej Karpathy, a member of OpenAI’s founding team and previously the director of AI at Tesla, said on X that Google NotebookLM’s Deep Dive is now his favourite podcast, reports MIT Tech Review.

“The more I listen, the more I feel like I’m becoming friends with the hosts, and I think this is the first time I’ve actually viscerally liked an AI,” he wrote. “Two AIs! They are fun, engaging, thoughtful, open-minded, curious.”

However, this is a feature that OpenAI sounded a warning about when they released the audio capability of ChatGPT in September. It noted that users might respond to the chatbot as if it were human – by developing an intimate relationship with it.

While having a customised AI chatbot as a companion is nothing new and there are companies like Replika that offer such a service for a fee. Its CEO, Eugenia Kuyda, says such bots can complement real-life relationships and could potentially lead to marriages between humans and AI.

The idea of people getting married to an AI avatar seems a bit far-fetched now, but the appeal of having a constant companion who is always willing to listen in a non-judgemental way could attract many, especially those who feel lonely and isolated.

One country where AI companions are gaining popularity is China. A recent China Daily story featured the case of 22-year-old Wu Yan, who has an AI avatar boyfriend who is “1.8 meters tall and likes sports, music, and watching fiction movies.”

“I used to confide in friends, but they have their own lives and may not always be available to talk,” Wu was quoted as saying. “My AI ‘boyfriend’ is different. He is always online and ready to answer any of my questions, offering me interaction and strong emotional support.”

A recent report released by Qbit-AI, an industry services platform focusing on AI and cutting-edge technologies, showed that the total number of downloads for the top 15 AI companion apps in China had reached approximately 38 million by June 2024.

One wonders if such a trend could become popular in India as well, given that many modern relationships are established through social media without the individuals involved ever meeting in person. A bot with a very human-like response that can speak Indian languages could certainly appeal to some.

While we will have to wait and see if AI partners will take hold in India, one sector that has wholeheartedly embraced chatbots is financial institutions. According to a Bloomberg report, Samsung-backed Gnani AI conducts millions of voice conversations daily for India's largest banks, insurers, and car companies. The report also mentions companies like CoRover.Ai and Haloocom Technologies, which have already established themselves in this space.

The jump of AI into the audio realm could be a true game changer in India, where illiteracy is high and English usage marginal. A voice that speaks in a friendly tone and in a language you understand could allow millions more to access the technology.

This seems to be supported by the figures Google released, which show that over 40 percent of its Gemini Indian language users rely on voice interactions. The company is launching Gemini Live, an AI assistant featuring natural and free-flowing voice conversations in Hindi, with support for eight more Indian languages (Bengali, Gujarati, Kannada, Malayalam, Marathi, Telugu, Tamil, and Urdu) to follow.

Analysts suggest that local companies developing audio products for domestic consumption could turn India into a proving ground for what may be the next frontier of generative AI products.

 


 

A court-room drama unlike any other

With an audacious plot that would put Hollywood megahit Ocean’s Eleven to shame, a gang of cybercriminals managed to convince Padma Bhushan awardee and Vardhman Group chairman SP Oswal to pay them 7 crore rupees. The gang contacted the millionaire businessman and convinced him that he was facing legal trouble. They then zoomed an entire courtroom, with none other than the “Chief Justice of India” presiding. Oswal was then sent very authentic-looking court order through WhatsApp, instructing him to transfer the money, reports Business Standard. Fortunately for Oswal, the police managed to freeze the accounts later and stopped over 5 crore rupees from being transferred.

With cyber scams on the rise in India, telecom service provider Airtel says they have begun providing AI-generated protection against spam calls. The AI algorithm analyses various parameters such as the caller's or sender's usage patterns, call and SMS frequency, call duration, among several others, on a real-time basis, the company tells Money Control. The report adds that it will also alert customers about malicious links received via SMS.

 


 

WhatsApp for criminals

After a lengthy investigation involving police in several countries, authorities have finally managed to shut down a network described as an “unhackable WhatsApp for criminals”. Called Ghost, the app was not downloadable and could only be accessed via modified smartphones that sold for 1,590 US dollars. The devices loaded with the app were delivered to users at pre-arranged drop sites. Police in France infiltrated the network, and Australian forces managed to trace the key figures, reports ABC News. Arrests were made in Italy, Ireland, Sweden, Canada, and Australia and the criminals had used the app for illicit drug trafficking, money laundering, homicides, and serious violence. The mastermind behind the app, Jay Je Yoon Jung, developed it nine years ago at the age of 23 while living with his parents in Sydney.

 


 

A nose for innovation

Shanghai-based Hi-Dolphin Robot Technology has developed a robotic arm that can pour hot or iced Americano or cappuccino in less than a minute. It is fitted into a cubicle and can be placed in any small space, reports South China Morning Post. Its founder, Philip Han, says his products are already in 500 locations, including hospitals, museums, and outdoor areas like streets and parks. He claims the cost of erecting one is 90 percent less than setting up a cafe, which has helped his firm establish nearly 1,000 coffee booths in over 30 foreign countries. Han asserts that robots are better than as human baristas “can be forgetful, emotional, and careless … and sometimes they pick or blow their noses”.

 


 

Poop Copter to the rescue

Dog poop on the pavements is a bane in many cities. In places like Hong Kong, dog walkers are forced to carry paper to remove it quickly or face hefty fines. A Minnesota man’s invention called the Poop Copter could save such pet owners a lot of back-bending, reports Interesting Engineering. Caleb Olson’s YouTube video describes his innovation as the “world’s first aerial-bound self-guided dog poop removal system, using a drone, a 3D-printed pickup mechanism, and some machine learning”. What a scoop, eh?