Artificial Intelligence

Robert J. Marks is one of the authors of Introduction to Evolutionary Informatics, with design theorist William Dembski and Winston Ewert. There’s little danger, he thinks, in computers ruling us but considerable danger that we can use them to magnify the impact of our errors. More. Here’s the podcast. See also: Human consciousness may not be computable One model of consciousness would mean that conscious computers are a physical impossibility. (Robert Marks)

Artificial intelligence crashes are historically common: First, what caused previous AI winters? There was one straightforward reason: The technology did not work. Expert systems weren’t experts. Language translators failed to translate. Even Watson, after winning Jeopardy, failed to provide useful answers in the real-world context of medicine. When technology fails, winters come. Nearly all of AI’s recent gains have been realized due to massive increases in data and computing power that enable old algorithms to suddenly become useful. For example, researchers first conceived neural networks—the core idea powering much machine learning and AI’s notable advances—in the late 1950s. The worries of an impending winter arise because we’re approaching the limits of what massive data combined with hordes of computers can Read More ›

Robert J. Marks looks at a recent theory: One issue to be aware of is this: AI is performed by computers and computers are entirely algorithmic. That is to say, they are constrained to obey a set of operations written by a computer programmer. Mathematics is algorithmically constructed, based on logic and foundational axioms. And physics is built algorithmically on foundational laws. In this sense, common naturalistic phenomena are largely algorithmic. They operate according to the logic of mathematics and the laws of physics. Penrose wondered, is anything in nature nonalgorithmic? He points to the collapse of a quantum mechanical wave function into a deterministic state. Such quantum effects can be found in the microtubules found in the brain. Penrose Read More ›

From Sir Martin Rees at NBC: AI apocalypse is certainly in the air. Elon Musk, Henry Kissinger, and the late Stephen Hawking have all predicted an AI doomsday. Industry professionals’ doubt and disparagement don’t seem to register with the media in the same way. Rees, who is former president of the Royal Society, goes further, however. He also predicts in his book that “a physics experiment could swallow up the entire universe.” When he received the Templeton Prize in 2011, he was noted for speculating that we could be living in a giant computer simulation. In 2017, he suggested that our universe may be lost in an “unbounded cosmic archipelago,” a multiverse where “we could all have avatars.” As for Read More ›

Readers may remember Dixon from the time MIT tried building a universal moral machine. Here are some of his thoughts on one overlooked aspect of the “superintelligent AI” myth: [Google AI researcher] Francois Chollet is right to recognize that we, like all animals, come pre-wired. Young deer stand, leap, and run within hours of birth. Birds build nests without prior instruction. Squirrels bury and find nuts. We speak and juggle abstract thoughts. But basic chemistry does not create language; while speaking may require chemical bonding and signaling, language rests on something more. Vision is another “chicken and egg” problem: The best human eye in the world is worthless without a nervous system to transmit the signals and a mind to interpret Read More ›

From Mind Matters: Much popular literature leaves the impression that living organisms are machines or even billions of them linked together. For example, at Medium, we learn, Brains receive input from the outside world, their neurons do something to that input, and create an output. That output may be a thought (I want curry for dinner); it may be an action (make curry); it may be a change in mood (yay curry!). Whatever the output, that “something” is a transformation of some form of input (a menu) to output (“chicken dansak, please”). And if we think of a brain as a device that transforms inputs to outputs then, inexorably, the computer becomes our analogy of choice… … But organisms differ Read More ›

Eric Holloway explains Jonathan Bartlett’s account of the human mind as a halting oracle: In his paper, “Using Turing oracles in cognitive models of problem-solving” Jonathan Bartlett proposes to model the human mind as a halting oracle. A brief explanation: Computer science pioneer Alan Turing (1912–1954) imagined a universal machine that can copy any other machine. However, this machine has a critical limitation: It cannot determine whether any given machine will run forever or not. This is known as the halting problem: “There can be no general procedure to decide if a self-contained computer program will eventually halt.” A halting oracle is a non-mechanical entity that can solve the halting problem for all machines. A common objection to Bartlett’s idea Read More ›

This post went viral yesterday at Mind Matters: The 2014 science fiction film Transcendence featured a scientist who uploaded his consciousness into an AI program. Many people talk as though things like that are just around the corner. But industry pros say it isn’t really possible. Why not? François Chollet, author of Keras, a framework for the Python deep learning language, offers a list of reasons, but starts by pointing to an underlying misconception: that a super-AI could be developed that would go on creating more super-AIs until something vastly more intelligent than a human being arises. He points out that such a process has not actually happened in the universe of which we have knowledge: An overwhelming amount of Read More ›

Researchers tried modeling intergroup anxiety but look what the public heard about the results, instead of the facts: In fact, nearly every claim about the paper seems to misunderstand how computer models work generally and how they worked in this paper in particular. First, there is nothing particularly “religious” about the criteria used in the model. In computer models, you can name the pieces of the model however you wish. The authors of the software simply happened to assign religious names to the components of the model. There was hardly anything religious about it apart from that. According to the BBC article, the study shows that “The most risky situations are when the difference in the size of two different Read More ›

In an effort to program self-driving cars to make decisions in a crisis, MIT’s Moral Machine offered 2.3 million people worldwide a chance to crowdsource who to kill and who to spare in a road mishap… The project aimed at building righteous self-driving cars revealed stark differences in global values. People from China and Japan were more likely to spare the old than the young. But in Western cultures, numbers matter more: The results showed that participants from individualistic cultures, like the UK and US, placed a stronger emphasis on sparing more lives given all the other choices—perhaps, in the authors’ views, because of the greater emphasis on the value of each individual. Karen Hao, “Should a self-driving car kill Read More ›

On the other side of a vast chasm… The code needed to detect and handle the flow between the situations increases polynomially with the number of driving situations we must address. That is, if we have 2 driving situations, there are 2 possible transitions to account for. If we have 3 driving situations, there are 6 possible transitions. If we have 4 driving situations, there are 12 possible transitions. Expressing it mathematically, for n driving situations, there are “n2 – n” transition possibilities. These types of numbers can mount up quickly. Therefore, every newly-identified driving scenario doesn’t just add one more scenario to code for in a linear fashion; it makes the project an order of magnitude more difficult. Many cheerleaders Read More ›

In his book, Life 3.0: Being Human in the Age of Artificial Intelligence, (2017), MIT physics prof Tegmark offers a science fiction scenario for how AI colossus Prometheus, produced by a group of idealistic programmers, the Omegas, could take over the world. And not only take over the world but make it a bureaucrat’s idea of a vastly better place. To top it off,  Prometheus produces an astonishing array of popular entertainment along the way: The Omegas noticed that after Prometheus had binge-watched a few hundred films, it started to get quite good at predicting what sort of reviews a movie would get and how it would appeal to different audiences. Indeed, it learned to write its own movie reviews in a Read More ›

The belief that there is no free will is a much more dangerous myth, he writes, at Mind Matters Today: There are four reasons to affirm the reality of free will against denial by materialist determinists. Two reasons are logical, and two are scientific. … 4. While scientific experiments do not entirely settle the matter, an objective review of the neuroscientific evidence unequivocally supports the existence of free will. The first neuroscientist to map the brains of conscious subjects, Wilder Penfield, noted that there is an immaterial power of volition in the human mind that he could not stimulate with electrodes. The pioneer in the neuroscience of free will was Benjamin Libet, who demonstrated clearly that, while there is an Read More ›

At Mind Matters Today, he explains, Along with millions of others, you are providing free training data: All of the most successful AI projects tend to follow a similar pattern. One of AI’s biggest needs is lots of data, and one of the most important tasks is finding ways to get people to provide them with the best data… for free. Currently, Facebook is utilizing hashtags applied to its Instagram photos to generate AI-based algorithms for detecting specific types of objects in images: Having so many images for training helped Facebook’s team set a new record on a test that challenges software to assign photos to 1,000 categories including cat, car wheel, and Christmas stocking. Facebook says that algorithms trained on Read More ›

Sloan Kettering molecular biologist argues that this may not be semantics. What if that’s what they are actually doing, in effect? One wonders, how would it affect cancer treatment? Abstract: Living organisms must maintain proper regulation including defense and healing. Life-threatening problems may be caused by pathogens or an organism’s own cells’ deficiency or hyperactivity, in cancer or auto-immunity. Life evolved solutions to these problems that can be conceptualized through the lens of information security, which is a well-developed field in computer science. Here I argue that taking an information security view of cell biology is not merely semantics, but useful to explain features of cell signaling and regulation. It also offers a conduit for cross-fertilization of advanced ideas from computer Read More ›