27 min

Is your AI talking back‪?‬ The CIO Australia Show

    • Technology

Earlier this month this was a good deal of excitement – probably mixed with horror for some – following ‘revelations’ in the tech and indeed mainstream press that one of Google’s chat bots had become sentient.
The call – or rather we might say alarm – was raised by software engineer, Blake Lemoine who asked a number of questions of the LaMDA chat bot he had been working with, which communicated several eerily-human responses including its desire to be considered a person with feelings and to be formally recognised as an employee at Google. 
LaMDA stands for the Language Model for Dialogue Applications and is a family of neural language models developed by Google. 
In internal memo to some 200 Google staff, Lemoine described a chat bot he’d been working with as a seven-year-old child that wants the world to be a better place for all.
Suffice to say Google issued a prompt response dispelling the idea that it was sentient, with Lemoine placed on forced leave.
Leaving aside all the predictable analogies with sci-fi books and films, the incident does raise interesting questions about the current capabilities of AI systems and what we might expect to see in coming years. 
We already have voice-driven systems – as well as chatbots – able to discern customer sentiment, so it’s not too farfetched to imagine more emotionally intuitive and responsive systems, say supporting sales and marketing, as well as health, in particular mental health.
Tune in to hear a fascinating conversation with Tathagat Banerjee, founder and CEO of Sydney-based startup Video Translator AI, as well as Julien Eps, head of the School of Engineering and Telecommunications with the University of NSW and a member of the NSW Smart Sensing Network.
As you’ll hear, while true sentience – itself a debatable term – is no doubt a stretch, Blake Lemoine’s experience serves to remind us that AI has nevertheless come a long way in a short period of time, with exciting new applications that effectively mimic human capabilities already in use.

Earlier this month this was a good deal of excitement – probably mixed with horror for some – following ‘revelations’ in the tech and indeed mainstream press that one of Google’s chat bots had become sentient.
The call – or rather we might say alarm – was raised by software engineer, Blake Lemoine who asked a number of questions of the LaMDA chat bot he had been working with, which communicated several eerily-human responses including its desire to be considered a person with feelings and to be formally recognised as an employee at Google. 
LaMDA stands for the Language Model for Dialogue Applications and is a family of neural language models developed by Google. 
In internal memo to some 200 Google staff, Lemoine described a chat bot he’d been working with as a seven-year-old child that wants the world to be a better place for all.
Suffice to say Google issued a prompt response dispelling the idea that it was sentient, with Lemoine placed on forced leave.
Leaving aside all the predictable analogies with sci-fi books and films, the incident does raise interesting questions about the current capabilities of AI systems and what we might expect to see in coming years. 
We already have voice-driven systems – as well as chatbots – able to discern customer sentiment, so it’s not too farfetched to imagine more emotionally intuitive and responsive systems, say supporting sales and marketing, as well as health, in particular mental health.
Tune in to hear a fascinating conversation with Tathagat Banerjee, founder and CEO of Sydney-based startup Video Translator AI, as well as Julien Eps, head of the School of Engineering and Telecommunications with the University of NSW and a member of the NSW Smart Sensing Network.
As you’ll hear, while true sentience – itself a debatable term – is no doubt a stretch, Blake Lemoine’s experience serves to remind us that AI has nevertheless come a long way in a short period of time, with exciting new applications that effectively mimic human capabilities already in use.

27 min

Top Podcasts In Technology

No Priors: Artificial Intelligence | Technology | Startups
Conviction | Pod People
Lex Fridman Podcast
Lex Fridman
All-In with Chamath, Jason, Sacks & Friedberg
All-In Podcast, LLC
Hard Fork
The New York Times
Acquired
Ben Gilbert and David Rosenthal
The Neuron: AI Explained
The Neuron