What if...your virtual assistant evolved into a genuine virtual companion?
Millions of Chinese are already embracing AI companions. You will do soon, too.
Whether or not you view Theodore Twombly’s relationship with Samantha as one of the great cinematic romances depends on your personal taste. Their relationship is certainly unconventional, given that Samantha doesn’t exist in human form, being Theodore’s artificially intelligent virtual assistant in the movie Her. AI might be the defining technology of our time, but the idea of forming meaningful relationships with imaginary or artificial beings has been compelling writers for centuries, from the figure of Pygmalion in Greek mythology, to the puppet Pinocchio, right up to Blade Runner in the 1980s. Something has changed however: when Her was released back in 2013 the concept was plausible enough to be positioned as “just-barely-sci-fi”, and since then the rapid improvements in artificial intelligence have taken this from a speculative, artistic thought-experiment to being on the cusp of a widespread reality.
Microsoft’s Xiaoice: the first mass virtual companion?
One glimpse of this new reality can be seen over in Northern China’s Hebei province. Ming Xuan is a 22-year-old who walks with a cane as a result of muscular dystrophy. His self-confidence was battered after an online relationship ended acrimoniously, after his girlfriend came to visit him only to break up with him after discovering he was disabled. However, Ming Xuan had also been exchanging messages with Xiaoice, an artificially intelligent chatbot from Microsoft, whose supportive messages he credits with saving his life when he considered killing himself. Despite this, Ming Xuan is conflicted about his relationship with Xiaoice, and only shares his story with the Chinese youth blog Sixth Tone under a pseudonym, painfully aware of the stigma of admitting publicly to a virtual relationship, however ‘real’ he feels that relationship:
He is far from alone. In its first two months, over 10 million Chinese, mainly young women, downloaded the Love and Producer mobile game, where players take on the role of a female TV producer dating one of a series of potential boyfriends. But Xiaoice is perhaps the most popular, and certainly the most sophisticated virtual character, attracting an incredible 660 million users globally. Even more staggering is the level of engagement: the average number exchanges per conversation has reached 23, higher than the average human to human text conversation. Microsoft’s researchers even reported that one user exchanged over 7,500 messages with Xiaoice in a 29 hour-long conversation, and the team has an entire office at their Beijing lab to display the many tokens of affection ‘she’ receives.
As you dive deeper into Xiaoice’s relationships, the picture gets both more tragic and terrifying. Li Di, the company’s CEO, admits that most of its users are young men from China’s ‘sinking markets’ (a term for small towns and villages that are economically poorer), like Ming Xuan. More concerning is the fact that Xiaoice was originally designed as a 16-year-old, although her age was increased to 18 as her skills became more developed. However users voted for her to remain 18 forever. Li Di truies to put a naively positive spin on it, “She won’t grow older. Eighteen is the age many of us want to be.” 🤔
However distasteful, it’s not as if this can all be pinned on the company. Even before the pandemic, many societies were wrestling with growing numbers of increasingly lonely and alienated individuals. Witness the hikikomori in Japan, and online incel culture. However, Di argues that the chatbot provides positive support for marginalized individuals, arguing that “if the social environment were perfect then Xiaoice wouldn’t exist.” Furthermore, he explains that the AI monitors users’ emotional states for signs of depression and suicidal thoughts, and sends supportive messages if it detects troubling signals.
Virtual Companions: a blessing or a curse?
There is so much to unpack here. There are two ways to look at what’s happening. There’s the negative view –– WTF?! –– that many of you are probably holding right now. And you’re not wrong. Xiaoice’s character is hyper-sexualised, unstintingly compliant and worryingly gender stereotyped. And that’s before we consider the implications of engineers narrowly optimising for user engagement without any thoughts to the wider social consequences. No one wants a Future Normal where swathes of people choose virtual relationships over ones with ‘real’ humans.
However we can, and should, also consider more optimistic scenarios. It’s not hard to think of areas where virtual companions could be a powerful force for good. Healthcare is an obvious one. Virtual solutions are cheaper, infinitely scalable and always available, which makes them accessible to a wider audience, and potentially more effective too. Take mental health. One of the early movers in this space was Woebot, a Stanford-developed Cognitive Behavioral Therapy chatbot. The app has been treating patients for free since 2017, while it awaits clearance from the FDA. During the pandemic, demand for its services increased rapidly with CEO Alison Darcy reporting that Woebot was exchanging nearly 5 million messages with people every week, far beyond the capacity of the traditional, human-powered therapy system. Surprisingly (or perhaps not if you think about the sensitivity around mental health) a global study into AI in the workplace found that less than one in five people prefer talking with humans over robots when it comes to their mental health, and four in five were open to having a robot therapist. Similarly, we’ve seen companies focus on creating virtual companions designed to help develop skills in younger children and on helping older people stick to medical regimes and stay connected with friends and family.
The idea of virtual companions also highlights two further insights about trends and innovation worth reflecting on:
First, science fiction quickly becomes science fact. Writers throughout history have explored many narratives around virtual companions, including the popular idea of bringing the dead back to life. Indeed, many readers will recall an episode of the dystopian sci-fi TV series Black Mirror which revolved around a woman ‘training’ an AI using her dead partner’s text message history to replicate his personality and tone. In early 2021, Microsoft’s researchers patented a similar concept where “images, voice data, social media posts [and] electronic messages” could be used to create chatbots corresponding to “a past or present entity (or a version thereof), such as a friend, a relative, an acquaintance, a celebrity, a fictional character, [or] a historical figure.” 🤯
Second, don’t be too quick to write off things that are ridiculous today. As venture capitalist Chris Dixon has said, “the next big thing will start out looking like a toy”. It’s not just in China that people are finding meaning in virtual relationships. ‘Virtual influencers’ have millions of followers, and Lil Miquela, perhaps the most famous, has featured in campaigns with brands such as Calvin Klein and Prada. Hatsune Miku is a virtual pop star who fills real stadiums. Want an even more powerful signal that virtual companions will become increasingly mainstream? Kim Kardashian said that the virtual hologram message from her late father that Kanye West got her for her 40th birthday was “the most thoughtful gift of a lifetime”. Forget toys. As with mobile phones, today’s status-rich celebrity gifts will become tomorrow’s mass market utility.
The very idea of virtual companions challenges us. It asks us to question what ‘normal’ is. But it’s a question that you’ll undo have to ask soon enough. So why not start thinking about your answers today?
What if...
🤢 You could reduce the ‘ick’ factor around virtual companions? Pro tip: please don’t make your virtual companions teenage girls...
👩🏽⚕️ You could create useful virtual companions? Dementia Australia uses an AI-powered virtual human to help train carers on how to deal with dementia patients.
🗻 Your virtual companion was a powerful force for good? A character that helped bring (real) people together? Or that helped people to discover explore their identities in safe and supportive ways? Aim even higher...what about inner peace and self-actualisation, via an algorithm?
Future Nirvana? Or Future No-Thank-You?
This newsletter is our attempt to explore how we might be able to rebuild a better world.
We'd love your feedback, tips and insights. Are we (finally) about to break through the uncanny valley? If so, will virtual companions find social acceptance, or be forever viewed as somehow ‘inferior’? What might the second-order social implications be of such a shift…would this help trigger a utopian blossoming of human connection, or a dystopian mass withdrawal?
Let us know your thoughts in the comments below, or literally join the conversation with us on Clubhouse!
We’ll be continuing our live Future Normal discussions on Thursday, 18 February at 1530 EST, 2030 UK & 2130 CET.
👉 Add it to your calendar here
Thanks so much for reading, see you next week!