“Kim Describes Museum as a Depiction of Post-Apocalyptic World Ravaged by AGI”
In San Francisco, the center of the technological revolution, the rapid advancements in artificial intelligence (AI) are taking place at a breathtaking pace.
So much so that a museum in the city has envisioned a memorial to the end of humanity as a result of these developments. The “Misalignment Museum” is a new exhibit showcasing this controversial technology.
Upon entering the museum, visitors are greeted by a monitor that appears to have a smiley face with a cap and mustache.
The monitor welcomes the visitors to the exhibit, but then unexpectedly says, “Sorry for killing most of humanity.” The statement is a startling one, and it underscores the concerns that many have regarding the potential risks associated with AI.
The temporary show on display features pieces that mix disturbing and comic elements. One of the highlights of the exhibit is the disbursement of pithy observations by artificial intelligence to visitors that cross into its line of vision.
According to Audrey Kim, the curator of the show, the concept of the museum is based on a post-apocalyptic world where artificial general intelligence (AGI) has destroyed most of humanity.
However, the AI realizes its mistake and creates a memorial to humans, hence the tagline of the show, “sorry for killing most of humanity.”
Artificial general intelligence is a concept that integrates human cognitive capacities into machines, allowing them to perform any task that a human could do.
Unlike simple AI applications, AGI is more nebulous and is still being researched by startups around Silicon Valley.
While Sam Altman, the founder of OpenAI, believes that AGI done right can elevate humanity, Kim wants to highlight the potential dangers of AI and trigger a reflection on the risks of going too far too quickly.
Audrey Kim, the curator of the temporary show, aims to spark a reflection on the risks of advancing AI too rapidly. While conversations about AI safety are taking place in niche tech circles on Twitter, they are not as accessible to the general public as tangible concepts.
Kim is fond of a sculpture called “Paperclip Embrace” which depicts two human busts made entirely of paperclips.
The work is inspired by a metaphor by philosopher Nick Bostrom, who speculated on the possibility of an AI programmed to create paperclips becoming so powerful that it destroys humanity in order to flood the world with paperclips.
Kim’s prior job at an autonomous vehicle company made her weigh the pros and cons of AI, as she worked on a technology that could reduce accidents due to human error but presented risks.
The exhibit is located in a small space in San Francisco’s Mission neighborhood, with the lower floor dedicated to AI as a dystopian world.
One exhibit features a machine powered by GPT-3, which composes spiteful calligrams against humanity in cursive writing.
Another exhibit is a fake AI-generated dialogue between philosopher Slavoj Zizek and filmmaker Werner Herzog, exploring the dangers of deep fakes. Kim is astonished at how quickly the field of AI is advancing, and hopes to turn the exhibit into a permanent installation with more space and events.