The web roundup for this month is a sequel to last month’s roundup on Mind, Consciousness, and Artificial Intelligence. I will address another interface between machines and minds, at the “hive mind” or the collective buzz of the internet, and the ways in which human limitations can be transmitted to the artificial minds that we beget. (And this one is also a bit late, though in this case it is attributable to human error, rather than to technical difficulties.)
Recently, Microsoft created Tay, an AI “chatbot” to appeal to 18- to 24-year-olds in the US. Tay was conceptualized as a teenage girl and existed entirely in Twitter, and Twitter was the source of Tay’s cumulative “intelligence.” Therein lies the problem. Less than a day later, Tay was tweeting racist epithets, denying the Holocaust, and calling game developer Zoe Quinn a “Stupid Whore.” Tay also apparently took a liking to Donald Trump.
In 1841, Ralph Waldo Emerson described the Over-Soul as something divine, which can serve as a check on the baser human natures: “Meantime within man is the soul of the whole; the wise silence; the universal beauty, to which every part and particle is equally related; the eternal ONE.” In 2016, the Internet appears to have formed a profane version of the collective unconscious, and this electronic over-soul strays far from Emerson’s lofty vision of transcendence. The Verge asks: “how are we going to teach AI using public data without incorporating the worst traits of humanity?” On Next Nature Network: “If the Internet was a person, then Tay would be a manifestation of a repressed and hateful part of its personality.” For Emerson, “Man is a stream whose source is hidden. Our being is descending into us from we know not whence.” For Tay, knowledge was also passed down from a hidden stream in the ether. For a while. Then Microsoft pulled the plug.
But perhaps there is still cause for optimism for the future of human-robot-human interactions? There is a robot monk, who offers spirituality in the form of a cute toy. People recoiled when they watched a Boston Dynamics employee push Atlas the robot over (ostensibly to demonstrate its successful bipedalism). Is a bit more difficult to interpret the implications of this research study: “The findings showed that when participants were instructed to touch the robot in areas that people usually do not touch, like the eyes or the buttocks, they were more emotionally aroused when compared to touching more accessible parts like the hands and neck. Participants also were more hesitant to touch these intimate parts based on the response times.” And then, of course, we are reminded: “From Siri to sexbots: Female AI reinforces a toxic desire for passive, agreeable and easily dominated women.”
The Weasel in the Machine
Though our image of the future may be cast in shades robotic, unemotional, and rational, the present reminds us that the world is still plenty capable of providing unexpected mishaps. A few days ago the Large Hadron Collider, at one of the most expensive and sophisticated scientific facilities in the world, was shut down by a weasel. The Collider has also been foiled when a bird dropped a baguette into the electrical system.
I have in the past pondered the ways in which scientists deploy probabilistic language to simultaneously disavow and bracket the chance of mishap; let’s recall: “in 2008, scientists even conceded that there was some risk of the Large Hadron Collider destroying the world, though it was considered extremely unlikely.” I can’t help but wonder: this is how the world ends? Not with a bang, but with a weasel?
Additional examples of how things don’t always go as planned:
- My Ph.D. in Crack
- CV of failures: Princeton professor publishes résumé of his career lows (download it here)