Your avatar awaits
An Auckland technology company is leading the way in creating more lifelike digital human helpers.
Something became obvious during the pandemic that served to undercut concerns about humans being displaced by artificial intelligence. As health workers, retail staff and call-centre agents all over the world fell sick with the virus, we faced a massive customer-service crunch. There weren’t enough nurses to screen patients for Covid symptoms in hospitals or tax agents to help laid-off workers claim income-relief payments.
Website chatbots, computer programs designed to simulate text-based conversations with a human user, became the default way to engage with banks or airlines swamped with frustrated customers. But chatbots are generally only useful for low-level queries and turn off customers seeking the human touch.
Auckland-based tech start-up Soul
Machines saw an opportunity to deploy its digital humans. Its AI-powered and fully rendered avatars appear on your computer screen, responding to your complaints with a concerned nod and a sensible answer. Using a web camera, some of them can even detect and respond to your facial expressions, making for a more natural conversation.
“As humans, we’re actually hardwired to engage face to face with each other, as
I’m discovering for myself again after two years of being locked in New Zealand,” says Greg Cross, Soul Machines’ chief executive and co-founder, who spoke to the Listener from San Francisco.
As the pandemic spread, Soul Machines partnered with the World Health Organisation to create a digital human that could answer people’s questions about the virus. Health care is a hot market for digital humans, but Soul Machines has customers across numerous sectors, from retail and entertainment to real estate and the automotive industry.
In February, the company raised US$70 million in venture capital funding, bringing the total invested so far in the six-year-old company to US$135 million. That likely makes it the most valuable New Zealand company working in the AI space. It employs more than 200 people
“I do see a world in the future where you and I will choose to create avatars to represent us in the Metaverse.”
and is rapidly adding staff in the US, its key market.
“Our specialist field of research is in the area of embodied cognition, which is different, for example, from what OpenAI is doing in natural language generation and DeepMind is doing in deep learning,” says Cross.
Those other companies are at the cutting edge of developing AI technologies that stand to make conversations with digital humans more free-flowing and natural. Where Soul Machines is breaking new ground is with its patented Digital Brain, which it uses to automate the animation of the digital avatars.
That technology has its roots in the work of Dr Mark Sagar, a two-time Oscar winner for science and engineering innovations who worked on Avatar and King Kong. Sagar pioneered a way of simulating how a human brain works – or at least, as much as we know about how the brain works.
“It means we can generate animation without having to, for example, do what the movie industry does, which is pre-record all the content using motion capture cameras,” Cross says. “It is delivered to you in the same way a Zoom call delivers real people into your device.”
Soul Machines could easily have its hands full creating digital humans for customer-service roles in the coming years. But its timing is also perfect to provide a key component of the next generation of the internet – known as the Metaverse.
“I do see a world in the future where you and I will choose to create avatars to represent us in the Metaverse, or what I simply call the 3D internet,” says Cross.
Animation techniques from the movie industry and video games can’t easily scale up to meet the demands of large virtual worlds, where dozens or hundreds of digital people may be interacting in real-time. Soul Machines’ Digital Brain is powering more realistic avatars, which many see as the key to us actually wanting to shop, learn, work and play in the Metaverse.
But Soul Machines will rely on advances in other areas of AI to make its digital humans more intuitive and lifelike. The current generation look convincing, but really only respond sensibly to a narrow range of queries. “The knowledge bases that they have access to, and the information that they use to converse with us, come from another field of AI, typically referred to as natural language processing,” Cross points out.
Curating that information requires AI systems to be trained, which is still a time-consuming process overseen by humans. But AI systems are also learning to train and even program themselves, speeding up the process of improving their repertoire and accuracy.
Cross points out that training and supporting new staff can be costly, so curating a knowledge base to power a digital human answering customer queries is a good investment, particularly in the face of ongoing labour shortages. “Where the Metaverse is today is about where the internet was in 1996,” Cross points out. “The dotcom bubble burst in 2001 and it was another six years before we saw the hardware device that drove the internet to scale — the first Apple iPhone.”
AI is still in search of its own iPhone moment, which will move us beyond the narrow abilities of widely used digital voice assistants such as Siri and Alexa to be more useful across a range of situations.
The rise of powerful, general-use AI will pose its own problems and “bad actors” will inevitably use it to exploit people, says Cross. “But, overwhelmingly over the course of time, the technology humans have created has enabled us to better the lives of the entire species. So I certainly remain an optimist.” die, in the hope they can be resurrected.
The Japanese robotics pioneer Hiroshi Ishiguro, who famously created an uncannily lifelike robotic replica of himself, believes there’s something unique about the Japanese that makes them more accepting of having robots in their lives.
Japan, he points out, is a small series of islands with a homogenous society that has been ruled by the same royal family
“If we don’t have AGI in the 21st century, something really bad will have happened to our species.”
for almost 2000 years. “We are like one big family and rather than forming a hierarchical society, we prefer to help each other find a role within it. Therefore, we aren’t concerned with distinguishing between robots and humans. We co-exist with one another,” he said in a Q&A at London’s Barbican Centre in 2019.
Ishiguro admits that advances in technology could make the differences between humans and robots irrelevant anyway. But Otsuki suggests the Japanese approach to developing humanoid robots puts a different spin on these rapidly evolving technologies. “It’s not about changing human beings so that they can be better, it’s about activating the potential that we have but might not be able to recognise or use without the help of technology,” he says.
Ultimately, he argues, technology is “human all the way down”. If we are going to have the ability to plug in neural implants and undergo life-enhancing gene therapies, it will still require people to make decisions at every step of the process, from how new genetic therapies are researched and trialled, to what information is used to train machinelearning systems to whether we will actually sign up for a Neuralink implant.
Otsuki considers New Zealand’s “precautionary” approach to developing new technologies to be similar to that of many European countries and more conservative than the permissive environments for innovation that exist in the US and China. “I don’t think one is necessarily better than the other. Do you want to move forward quickly, but maybe have to deal with a whole lot of unintended consequences and harms that nobody expected? Or do you want
regulators and the government to take a more careful approach?”
STRONGER GUARDRAILS
That’s the dilemma politicians face in the transhuman age, although with few of them able to truly understand the technology and its potential to change society, they too often treat it as a curious sideshow.
That’s what led to the dominance of Big Tech in our lives. Lax regulation and an infatuation with the brilliance of Silicon Valley entrepreneurs allowed the likes of Google and Facebook to develop their digital platforms largely unhindered, letting them to scoop up masses of behavioural data about billions of people. The resulting disruption to entire industries, and even the democratic process, has been profound – and not always in a good way.
Unintended consequences in the form of misinformation, hate speech, monopolistic practices and worsening inequality have regulators scrambling to catch up with the innovators. “Facebook had the opportunity to change their algorithms, to incite less harm,” Otsuki says. “They chose not to do anything about it. They made the decision to rile people up because that helped engagement.”
With the next wave of technologies able to literally get inside our heads and manipulate our biology, the guardrails may need to be much stronger, Otsuki argues. At the very least, we will need a greater level of accountability and transparency on the part of the companies and entrepreneurs commercialising these futuristic technologies.
Otsuki is sceptical of letting the benevolent tech oligarchs get on with it on our behalf. “I would not want Elon Musk running things. He might know technology well, but I think he probably knows very little about society or has a very narrow view of what society is.”
Musk’s attempt to buy Twitter for US$44 billion, delisting it from the New York Stock Exchange in the process, Otsuki says, “speaks to his mindset. Some of his moves feel like he’s trying to avoid accountability for making big decisions.”
In the early 19th century, angry English textile workers rampaged through cotton factories, smashing the mechanised looms that had been introduced. The technology represented a threat to their artisanal ways of working. But while they turned their hammers on the new looms, their anger was really directed at the mill owners aiming to hire unskilled workers to run the machines – and churn out what the textile workers considered to be inferior products.
The loom smashing became so disruptive that Parliament passed a law imposing the death penalty for those guilty of the destruction and sent troops into mills to maintain order. “The Luddites only started sabotaging the machines when they weren’t allowed to organise, they weren’t allowed to have a voice through other means,” says Otsuki. “If they’d had a way to speak back to the people putting those technologies in place, there might have been a different way forward.”
NO ENDGAME
There’s a lesson there for the transhuman age. Today, someone who eschews owning a mobile phone or doesn’t use a computer is considered a Luddite. But with millions of jobs likely to be displaced due to robotics and AI-powered automation in the coming decades, it’s only natural that new technologies will face similar opposition.
“Some people are going to lose out here for a time, but overall, things are going to get much better,” Elise Bohan maintains.
Technologists should be honest about the downsides of augmenting and improving humans through technology, but putting a handbrake on progress is futile. Social acceptance of transhumanism will come about as people realise the technology makes for easier living.
Bohan thinks that will particularly apply to advances in biotechnology that reduce disease, ageing and some of the more uncomfortable processes of being human. Artificial wombs, which have already been used to nurture and grow baby lambs outside of their mothers’ wombs, would be like the “iPhone of biotechnology”, Bohan says. “You are getting rid of the thing women absolutely hate about having children, which is being a walking incubator for nine months.”
The same goes for the holy grail of human health, which is life-extension technologies. “When you ask people whether they want to live to 150, they say no, they’d be terribly bored,” says Bohan. “But if I can age at a slower rate with more good years and be sprightly for longer, people want that.”
The most radical technologies may not arrive in a rapid rush, but gradually, just as mobile phones and social media crept up on us and, over a 20-year period, became part of the modern world. Transhumanists see this as just the natural evolutionary progression of the human race. Our intelligence has brought us to a point in human history where we can write our own biological future.
“It feels to us like such a miraculous thing, that these tiny single-celled organisms go on to build these giant cities, spaceships and light up the face of the planet,” says Bohan. “It just seems so obvious that more knowledge and more discovery and more complexity, more joy, more experience, would be a great thing.”
But what is it all in aid of ? Where is the evolutionary road map taking us? “The only intent of evolution has ever been survival and reproduction. I think that’s very frightening for people because we want a sense of anthropomorphised meaning. We want a sense of spirituality and greater cosmic significance.”
There’s no obvious endgame for this antientropic force that drives species to adapt to their changing environment. “If there’s no objective value to that, so be it,” says Bohan. “But there is subjective value to that and intelligent life can do cool things. Maybe, just maybe, there is some higher meaning somewhere that if we were smart enough, we could figure it out. That’d be cool, too.” ▮
“If [you] can age at a slower rate with more good years and be sprightly for longer – people want that.”
Future Superhuman: Our transhuman lives in a make-or-break century, Elise Bohan, Hachette Aotearoa, $34.99