A museum exhibit has imagined a post-apocalyptic world after artificial intelligence (AI) has wiped out most of humanity, before attempting to make amends by creating a museum exhibit.
Called the Misalignment Museum, the team hopes the exhibit will increase visitors' knowledge and awareness of artificial general intelligence (AGI) – the ability of an AI to think and perform tasks like a human.
"Our goal is to created a space to reflect on the technology itself and think critically about Artificial Intelligence and its implications," the team writes on their website. "Our hope is to inspire and build support to formulate and enact risk mitigation measures we can take to ensure a positive future as AI continues to develop, and in the advent of AGI."
The exhibit includes a friendly welcome from the machines, reading "sorry for killing most of humanity". The exhibit includes a number of installations, from Michelangelo's The Creation of Adam as categorized by AI, to a team of cans of spam powered by language model trained on a "piggy" version of Aldous Huxley's Brave New World attempting to tell their own story.
The exhibit also features a fairly convincing fake conversation between an AI-generated Werner Herzog and philosopher Slavoj Žižek
While innovations in natural language processing may be impressive, they work more like "spicy autocomplete" than actual general artificial intelligence.
Some estimates have placed artificial general intelligence as being as little as seven years away, though a survey of AI researchers put it more around 2050.
But will AGI necessarily be as bad as the apocalyptic exhibit makes out? On that, AI researchers have their concerns.
Surveying researchers who had co-authored at least two computational linguistics publications between 2019–2022, a team found that people in the field are not altogether positive about the direction of travel for AI.
"73 percent of respondents agree that labor automation from AI could plausibly lead to revolutionary societal change in this century, on at least the scale of the Industrial Revolution," the researchers wrote of their survey. Meanwhile, a surprisingly-high 36 percent of AI researchers agreed that AI could produce catastrophic outcomes this century "on the level of all-out nuclear war".
It's not even clear that in real life the AI would apologize with a museum afterward. The Misalignment Museum exhibit is currently open to visitors in San Francisco.