Photo: Empirics Asia
Cover Empirics Asia founder Melvin Poh and his team just launched the Empirics Podcast, a show produced entirely by artificial intelligence. (Photo: Empirics Asia)

Thoughts from Empirics Asia's founder, whose team just released the world's first podcast produced and run entirely by a machine learning artificial intelligence

"Listening to an AI talk about art is really interesting," quips lawyer, entrepreneur and investor Melvin Poh, founder of open access knowledge sharing platform Empirics Asia. "Definitely something to cross off on the bucket list in 2021."

If you too are curious to hear an artificial intelligence (AI) discuss the science of happiness, mental health in a pandemic or the 'freemium' economy, the Empirics Podcast is one for your playlist. Launched this year, the Empirics Podcast is the world's first podcast show prepared, produced and hosted entirely without human intervention by a proprietary machine learning AI developed by Empirics Asia.

Related: 6 Malaysian Financiers Predict The New Industries That Will Impact The Future

Tatler Asia
Photo: Empirics Asia
Above Photo: Empirics Asia

As a student at Harvard University years ago, Poh founded The Asian Entrepreneur, a business publication that eventually expanded into the crowdsourced knowledge portal known today as Empirics Asia. The AI behind Empirics Podcast was initially built and trained to filter and manage crowdsourced articles and insights from over a thousand contributors around the world on the Empirics Asia platform. 

More: 10 Cool Malaysian Podcasts To Stay Connected In 2021

Six months ago, Poh and his team acquired the necessary funding to develop it further; training an AI already familiar with the platform's curated content to take the next step and create short podcast episodes with the help of sophisticated narration software.

But can AI truly curate audio content meaningfully? Is there validity to the fear that AI could one day replace humans in content creation? Poh sheds light on these questions in the following Q&A.

What's your take on the negative stereotypes of AI today?

There are very real fears people have about losing their jobs to AI. The fear of automation replacing jobs has been around throughout history. But truthfully, AI has the potential to create new jobs in the economy. We might see economies becoming overall more efficient and potentially new industries being created because of it.    

What first gave you the idea to use artificial intelligence in your organisation?

At Empirics Asia, we wanted to find a way to manage the editorial process more efficiently. We had so much content coming in without knowing what was relevant to readers at a given time. So, we started to look into how we could integrate artificial intelligence into our editorial process, building something that would help edit and filter the data coming in while revealing what knowledge fields we should focus on publishing. We soon realised that we had the tools to do something more creative. That's how this podcast came about. 

More: Award-Winning Malaysian Scientist Dr Serena Nik-Zainal Speaks Up About Breaking The Glass Ceiling

Tatler Asia
Photo: Empirics Asia
Above Photo: Empirics Asia

Did you encounter any hiccups with the first few episodes? 

Due to the parameters we designed for it, this technology only uses information from our platform, which is already filtered, so in terms of content, it's relatively safe. The issues we’ve experienced had more to do with delivery.  

For instance, in one of the episodes on cryptocurrency, it used the word ‘foolish’ to describe the Turkish government's regulation of cryptocurrency. When I heard that, my jaws dropped (laughs). On this platform we want to remain neutral, and that's something we constantly have to monitor. 

We removed that part and introduced a parameter into it so it would learn over time not to repeat that mistake. It's funny because it's very much like teaching a child. There's a lot we still have to work out and a lot of times, we can’t guess what will happen next.  

How does this podcast help tackle negative perceptions about AI? 

Besides the fear of AI, there’s also this sort of discrimination against it. Deep down people might think, "Why should I listen to a robot?" Because it can deliver value and knowledge in more efficient and more accurate ways than we ever could as humans. Narration technology is improving so rapidly. We made sure to create a voice profile that’s nice to listen to and naturalistic. Hopefully it will help acclimatise and educate more people about the value of AI.   

Related: Xiaomi's K.M. Leong On What Puts The Mi 11 Series In A League Of Its Own

Given the podcast's overall success, how can this technology be applied in more sectors in Malaysia?

Education, now more than ever, is going through massive disruptions. We're seeing it become more digital. If educators were to integrate this technology into their infrastructure, it could potentially reduce cost of education significantly by reducing the cost of delivering relevant information. In future, you could potentially go to Ivy League schools around the world without stepping foot there. Imagine having a Harvard-curated class delivered by a Harvard-managed AI that's tapped into the school's knowledge and resources.  

What's your next step for this machine-learning AI?

The next possible area would be video creation. The technology is still limited and so far, it is impossible for an AI to produce the kind of video content we want to achieve. But that's something we're exploring now. 

Related: Astrophysicist Datuk Mazlan Othman Reflects On A Life Dedicated To Science 

 

Topics