Various versions of the Turing test have been put forward over the years but only one is so tough that even humans haven’t yet passed it. That will change if Florentin Neumann at the University of Paderborn in Germany and a couple of pals have their way.
This alternate exam is called the Hofstadter-Turing Test, after Douglas Hofstadter who put forward a version of the idea in an essay called Coffee House Conversation in 1982. Here’s how it works (pay attention because it contains a certain circularity to the argument):
An entity passes the Hofstadter-Turing Test if it first creates a virtual reality, then creates a computer program within that reality which must finally recognise itself as an entity within this virtual environment by passing the Hofstadter-Turing Test.
Spot the tricky circularity to this test? Players can only pass if they create a virtual intelligence which must then pass the test itself. And since that hasn’t been achieved by any human in history, nobody has yet passed.
What’s interesting about the paper though, is that Neumann and co claim that humanity is moving closer to achieving a pass. First of all, we’re half way there because we’ve already built various virtual worlds. And now Neumann and co claim to have implemented a version of the Hofstadter-Turing Test in the Second Life virtual world.
“We have succeeded in implementing within Second Life the following virtual scenario: a keyboard, a projector, and a display screen. An avatar may use the keyboard to start and play a variant of game classic Pac-Man, i.e. control its movements via arrow keys.”
They go on to say:
“With some generosity, this may be considered as 2.5 levels of the Hofstadter-Turing Test:
1st: The human user installs Second Life on his computer and sets up an avatar.
2nd: The avatar implements the game of Pac-Man within Second Life.
3rd: Ghosts run through the mace on the virtual screen.
Observe that the ghosts indeed contain some (although admittedly very limited)
form of intelligence represented by a simple strategy to pursue Pacman.”
They’re absolutely right that taking this on board requires a remarkable amount of generosity: the Ghosts in a Pacman game are unlikely to ever put in a decent challenge in any other type of Turing Test.
But suppose we give them the generosity they desire. The process raises some interesting ways of analysing the various levels of reality that could occur when machines become intelligent. And what of the possibility that our efforts may be validating the intelligence of a programmer exactly one level higher than us?
Ref: arxiv.org/abs/0904.3612: Variations of the Turing Test in the Age of Internet and Virtual Reality
This new data poisoning tool lets artists fight back against generative AI
The tool, called Nightshade, messes up training data in ways that could cause serious damage to image-generating AI models.
Rogue superintelligence and merging with machines: Inside the mind of OpenAI’s chief scientist
An exclusive conversation with Ilya Sutskever on his fears for the future of AI and why they’ve made him change the focus of his life’s work.
Data analytics reveal real business value
Sophisticated analytics tools mine insights from data, optimizing operational processes across the enterprise.
Driving companywide efficiencies with AI
Advanced AI and ML capabilities revolutionize how administrative and operations tasks are done.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.