These are the rich ideas behind FRIEND, a new exhibition opening at the National Communication Museum on 1 November that flips the usual narrative about artificial intelligence and automation on its head. Instead of focusing on job displacement or existential threats, the show asks a more intimate question: what happens when we design machines to be vulnerable, affectionate, and imperfect?
The six-month exhibition runs until 26 April 2026 in the museum's Hawthorn galleries, bringing together robotics research from Japan, Australia and beyond. It's a collaboration between the National Communication Museum and studioBOWL that arrives at a cultural moment when AI dominates headlines, usually in apocalyptic terms. FRIEND offers something different - a chance to actually spend time with robots rather than just worry about them.
Why weak robots matter more than strong ones
Let's start with the Weak Robots from Japan, which represent a radical departure from traditional robotics philosophy. For decades, the field has pursued strength, precision and reliability - machines that perform tasks better than humans ever could. Weak Robots deliberately fail at these goals. They stumble. They forget what they're doing. They need help from people nearby to complete simple tasks.
This approach emerged from Japanese research into social robotics and human-robot interaction. Rather than viewing failure as a design flaw to be eliminated, researchers began exploring whether vulnerability could become a feature that strengthens human-machine bonds. The hypothesis: people form deeper connections with things that need them.
The principle has precedents in human psychology. The Benjamin Franklin effect describes how doing favours for someone makes us like them more, not less. Weak Robots apply this to human-machine relationships, suggesting that robots who need human assistance might integrate into social environments more successfully than hyper-competent machines that operate independently.
By now, this might be freaking you out even more.
But this specifically matters because companion robots are increasingly marketed for elderly care, disability support and mental health applications. A perfectly efficient care robot might complete all necessary tasks while leaving the person feeling obsolete or infantilised. A robot that occasionally needs help, that shows vulnerability, might preserve the person's sense of agency and purpose.
The question becomes: what kind of dependence do we want to create?
The fundamental question behind chatbots
ELIZA, the 1960s therapy chatbot revived for FRIEND, reveals how little the fundamental questions have changed. Created by MIT computer scientist Joseph Weizenbaum, ELIZA used simple pattern matching and scripted responses to simulate a Rogerian psychotherapist (the supportive, non-judgmental psychologist that allows you to discover solutions for yourself).
The program would take statements from users and reflect them back as questions, creating an illusion of understanding without any actual comprehension.
Weizenbaum became disturbed by how quickly people attributed real intelligence and empathy to ELIZA, even when they knew it was just a program. His secretary asked him to leave the room so she could talk to ELIZA privately. Psychiatrists proposed using it for actual therapy. The response alarmed Weizenbaum enough that he spent the rest of his career warning about artificial intelligence and the dangers of replacing human relationships with computational ones.
Six decades later, AI chatbots are everywhere - customer service, mental health apps, companion programs. The technology is vastly more sophisticated than ELIZA, using large language models and natural language processing that Weizenbaum couldn't have imagined.
But the fundamental question remains: why are humans so willing to perceive intelligence, understanding and care in machines that don't actually possess these qualities?
Elena Knox's Protective Seal installation takes this question to an extreme environment. The work documents a Japanese therapy robot, originally designed for dementia care and emotional support, being taken to the Arctic.
Surrounded by ice and emptiness, the seal robot becomes a strange ambassador of designed comfort in a landscape that predates and will outlast all human technology. The video installation captures something uncanny about care machines - they continue their programmed behaviors regardless of context, offering the same gentle movements and responsive sounds whether in a nursing home or at the edge of the world.
The National Communication Museum is a must-visit experience
The National Communication Museum opened just over a year ago in a refurbished 1930s telephone exchange, bringing to life a heritage telecommunications collection that volunteers have been gathering since the 1960s. But the museum isn't just displaying historical objects - it's actively building its own robotics capability through NCM Studio, the in-house team responsible for reviving, modifying and creating the technological works throughout the building.
For FRIEND, the studio team rewired Furbies to create a responsive choir, revived ELIZA on period-appropriate hardware running on emulation mainframes, and brought the discontinued Pepper robot back into operation. This hands-on approach to exhibition-making reflects a broader philosophy: understanding technology requires getting inside it, repairing it, modifying it, and sometimes breaking it.
The Furby Choir demonstrates this approach. Some of the toys were loaned from the National Communication Museum community, carrying the wear and affection of their original owners. Others were sourced and then hacked by the studio team, their internal circuits modified so they could respond collectively when visitors ask what is a friend. The resulting installation blends nostalgia for 1990s toy culture with questions about how consumer electronics create emotional attachments through relatively simple feedback loops - sounds, movements, blinking eyes.
Needless to say, the Museum is a totally unique and rather essential experience in Melbourne.
Opening weekend brings global robotics research to Melbourne
The 1 November opening program extends the exhibition's reach beyond the gallery space with keynote lectures, panel discussions and live demonstrations bringing international researchers to Melbourne. Professor Michio Okada from Toyohashi University of Technology's Interaction & Communication Design Lab headlines the programming, presenting research on how robots can facilitate human connection rather than replace it.
Okada's lab developed several of the Weak Robots in the exhibition and has published extensively on social robotics, communication design and human-robot interaction. His research challenges assumptions that robots should be as humanlike as possible, instead exploring how limited, specialised or even frustrating machine behaviours might serve social functions that perfect automation cannot.
The day concludes with a demonstration of JIZAI ARMS, wearable robotic limbs developed through collaboration between Professor Masahiko Inami's JIZAI Body Project and Professor Shunji Yamanaka's Prototyping & Design Laboratory at the University of Tokyo. The arms explore human augmentation - treating technology as extension of the body rather than external tool. One audience member will try the system, experiencing how additional robotic limbs affect proprioception, spatial awareness and physical capability.
Perhaps this is the final question the exhibition leaves us with. Rather than asking how humans and machines can coexist as separate entities, augmentation technology asks how machines might become part of human bodies and identities.
The implications span disability support, labor practices, sports, military applications and elective body modification.
FRIEND runs through April 2026. For more information, head here.