Pepper, the world’s smartest retail robot, is changing how we think about AI

If you’ve been fortunate enough to jet off to Japan recently, you may have checked in to a hotel only to be greeted by a velociraptor robot wearing a concierge hat and a bowtie.    

While this may seem like a cheesy gimmick, in countries such as Japan this automation of services indicates the direction in which artificial intelligence (AI) is heading, as machines become more advanced and capable of carrying out human tasks.   

But with that comes key questions, and concerns, about what the future of AI looks like, and the ever-increasing role these machines will play in sectors spanning retail and health, to aged care.  

Enter Pepper, one of the world’s first social humanoid robots who has been specifically designed for human interaction. Pepper, who has a human-like demeanour and a warm smile, is already proving they are not just a pretty face, having been deployed across more than 2,000 workplaces globally.  

Not only can this bot dance and cheer (a 100-strong team of Peppers nabbed themselves a Guinness World Record for the largest robot cheering squad) but they can also recognise people’s faces, make gestures and even hold a conversation. 

Now Canberrans will have a chance to get up-close-and-personal with Pepper, who has found a new (temporary) home at the Kambri Cultural Centre at The Australian National University (ANU). 

It’s all part of the Uncharted Territory arts and innovation festival, an initiative led by the ACT Government with support from ANU.  

“The public will have a chance to engage with Pepper and ask questions, and for many people this will be the first time they’ve had the chance to interact with an intelligent robotic being,” Associate Professor Damith Herath, from the University of Canberra (UC), says.  

Herath, who leads UC’s Collaborative Robotics Laboratory, has dedicated his career to developing more intelligent bots, with the goal of finding the perfect way robots and humans can coexist and work in harmony. 

“We know that some people find robots, and AI more generally, quite intimidating and so we really want to try and debunk some of the common misconceptions and preconceived notions they may have,” he says.  

Powered by the language processing tool, ChatGPT, Pepper has access to a wealth of information at their fingertips. The intelligent bot has facial recognition technology and an array of microphones and speakers to help them interpret speech and process information in real-time. 

“We’re certainly a long way off seeing an uprising.”

Associate Professor Jenny Davis

Putting Pepper in a challenging and dynamic environment, where the robot can have conversations with multiple people at once, is about more than just flaunting the machine’s technological prowess, ANU Associate Professor Jenny Davis says. 

“We’ll be looking at the extent and type of emotional and social connections participants develop with Pepper,” Davis, Deputy Director of the ANU Humanising Machine Intelligence project, says. 

“Before and after the interactions we’ll be handing out questionaries to quantify participants’ experiences and gauge how effectively they feel Pepper understood them.” 

The lessons learnt from this experiment will help inform the practical application of deploying robots like Pepper in the real world, particularly in sectors such as aged care where emotional connections and empathy are crucial. 

“How you integrate the use of robots to assist in disaster response missions compared to rolling them out in aged care homes are two completely different situations,” Davis says.  

“You probably don’t want the people who are working in emergency or rapid response situations to form a strong emotional or cognitive connection to the robot, because it’s most likely going to get destroyed.  

“But if you’re deploying a robot in an aged care setting, it can be really important for the residents who are encountering that robot to feel comfortable and at ease around it, so we need to consider how we adjust those design decisions depending on the setting.” 

While Davis says there are clear benefits to the use of AI in our everyday lives and in the workforce, there are still hurdles that developers, employers and policymakers must address.

This includes overcoming AI bias.

“For instance, in an employment situation we know that hiring algorithms have consistently favoured male candidates, and often white candidates, and the reason for that is because those algorithms use data from that company’s previous employment practices,” Davis says.  

“Thinking about how we can reconfigure or remodel these systems to recognise and overcome those biases so you have a more equal outcome is really a key question that is driving much of the discussion around ethics in AI today.” 

The technological capabilities of Pepper are certainly impressive, and a testament to how far advancements in AI have come, but both Damith and Davis say we shouldn’t be too worried about robots taking over humanity any time soon.  

“I don’t think pausing developments in AI is plausible or very effective, and is not the way we should go. Creating an infrastructure of slower and more considered development is probably a more reasonable and viable way to think about resolving many of the issues and risks that surround AI,” Davis says.  

“It’s up to all of us to take responsibly and to make sure progressions in AI technologies are managed safely and ethically. Although one thing is for sure, we’re certainly a long way off from seeing an uprising by Skynet.” 

Thank goodness for that, because we’ve seen Terminator, and don’t like how that turned out.