Tag Archives: laws
#437982 Superintelligent AI May Be Impossible to ...
It may be theoretically impossible for humans to control a superintelligent AI, a new study finds. Worse still, the research also quashes any hope for detecting such an unstoppable AI when it’s on the verge of being created.
Slightly less grim is the timetable. By at least one estimate, many decades lie ahead before any such existential computational reckoning could be in the cards for humanity.
Alongside news of AI besting humans at games such as chess, Go and Jeopardy have come fears that superintelligent machines smarter than the best human minds might one day run amok. “The question about whether superintelligence could be controlled if created is quite old,” says study lead author Manuel Alfonseca, a computer scientist at the Autonomous University of Madrid. “It goes back at least to Asimov’s First Law of Robotics, in the 1940s.”
The Three Laws of Robotics, first introduced in Isaac Asimov's 1942 short story “Runaround,” are as follows:
A robot may not injure a human being or, through inaction, allow a human being to come to harm.
A robot must obey the orders given it by human beings except where such orders would conflict with the First Law.
A robot must protect its own existence as long as such protection does not conflict with the First or Second Laws.
In 2014, philosopher Nick Bostrom, director of the Future of Humanity Institute at the University of Oxford, not only explored ways in which a superintelligent AI could destroy us but also investigated potential control strategies for such a machine—and the reasons they might not work.
Bostrom outlined two possible types of solutions of this “control problem.” One is to control what the AI can do, such as keeping it from connecting to the Internet, and the other is to control what it wants to do, such as teaching it rules and values so it would act in the best interests of humanity. The problem with the former is that Bostrom thought a supersmart machine could probably break free from any bonds we could make. With the latter, he essentially feared that humans might not be smart enough to train a superintelligent AI.
Now Alfonseca and his colleagues suggest it may be impossible to control a superintelligent AI, due to fundamental limits inherent to computing itself. They detailed their findings this month in the Journal of Artificial Intelligence Research.
The researchers suggested that any algorithm that sought to ensure a superintelligent AI cannot harm people had to first simulate the machine’s behavior to predict the potential consequences of its actions. This containment algorithm then would need to halt the supersmart machine if it might indeed do harm.
However, the scientists said it was impossible for any containment algorithm to simulate the AI’s behavior and predict with absolute certainty whether its actions might lead to harm. The algorithm could fail to correctly simulate the AI’s behavior or accurately predict the consequences of the AI’s actions and not recognize such failures.
“Asimov’s first law of robotics has been proved to be incomputable,” Alfonseca says, “and therefore unfeasible.”
We may not even know if we have created a superintelligent machine, the researchers say. This is a consequence of Rice’s theorem, which essentially states that one cannot in general figure anything out about what a computer program might output just by looking at the program, Alfonseca explains.
On the other hand, there’s no need to spruce up the guest room for our future robot overlords quite yet. Three important caveats to the research still leave plenty of uncertainty to the group’s predictions.
First, Alfonseca estimates AI’s moment of truth remains, he says, “At least two centuries in the future.”
Second, he says researchers do not know if so-called artificial general intelligence, also known as strong AI, is theoretically even feasible. “That is, a machine as intelligent as we are in an ample variety of fields,” Alfonseca explains.
Last, Alfonseca says, “We have not proved that superintelligences can never be controlled—only that they can’t always be controlled.”
Although it may not be possible to control a superintelligent artificial general intelligence, it should be possible to control a superintelligent narrow AI—one specialized for certain functions instead of being capable of a broad range of tasks like humans. “We already have superintelligences of this type,” Alfonseca says. “For instance, we have machines that can compute mathematics much faster than we can. This is [narrow] superintelligence, isn’t it?” Continue reading
#437816 As Algorithms Take Over More of the ...
Algorithms play an increasingly prominent part in our lives, governing everything from the news we see to the products we buy. As they proliferate, experts say, we need to make sure they don’t collude against us in damaging ways.
Fears of malevolent artificial intelligence plotting humanity’s downfall are a staple of science fiction. But there are plenty of nearer-term situations in which relatively dumb algorithms could do serious harm unintentionally, particularly when they are interlocked in complex networks of relationships.
In the economic sphere a high proportion of decision-making is already being offloaded to machines, and there have been warning signs of where that could lead if we’re not careful. The 2010 “Flash Crash,” where algorithmic traders helped wipe nearly $1 trillion off the stock market in minutes, is a textbook example, and widespread use of automated trading software has been blamed for the increasing fragility of markets.
But another important place where algorithms could undermine our economic system is in price-setting. Competitive markets are essential for the smooth functioning of the capitalist system that underpins Western society, which is why countries like the US have strict anti-trust laws that prevent companies from creating monopolies or colluding to build cartels that artificially inflate prices.
These regulations were built for an era when pricing decisions could always be traced back to a human, though. As self-adapting pricing algorithms increasingly decide the value of products and commodities, those laws are starting to look unfit for purpose, say the authors of a paper in Science.
Using algorithms to quickly adjust prices in a dynamic market is not a new idea—airlines have been using them for decades—but previously these algorithms operated based on rules that were hard-coded into them by programmers.
Today the pricing algorithms that underpin many marketplaces, especially online ones, rely on machine learning instead. After being set an overarching goal like maximizing profit, they develop their own strategies based on experience of the market, often with little human oversight. The most advanced also use forms of AI whose workings are opaque even if humans wanted to peer inside.
In addition, the public nature of online markets means that competitors’ prices are available in real time. It’s well-documented that major retailers like Amazon and Walmart are engaged in a never-ending bot war, using automated software to constantly snoop on their rivals’ pricing and inventory.
This combination of factors sets the stage perfectly for AI-powered pricing algorithms to adopt collusive pricing strategies, say the authors. If given free reign to develop their own strategies, multiple pricing algorithms with real-time access to each other’s prices could quickly learn that cooperating with each other is the best way to maximize profits.
The authors note that researchers have already found evidence that pricing algorithms will spontaneously develop collusive strategies in computer-simulated markets, and a recent study found evidence that suggests pricing algorithms may be colluding in Germany’s retail gasoline market. And that’s a problem, because today’s anti-trust laws are ill-suited to prosecuting this kind of behavior.
Collusion among humans typically involves companies communicating with each other to agree on a strategy that pushes prices above the true market value. They then develop rules to determine how they maintain this markup in a dynamic market that also incorporates the threat of retaliatory pricing to spark a price war if another cartel member tries to undercut the agreed pricing strategy.
Because of the complexity of working out whether specific pricing strategies or prices are the result of collusion, prosecutions have instead relied on communication between companies to establish guilt. That’s a problem because algorithms don’t need to communicate to collude, and as a result there are few legal mechanisms to prosecute this kind of collusion.
That means legal scholars, computer scientists, economists, and policymakers must come together to find new ways to uncover, prohibit, and prosecute the collusive rules that underpin this behavior, say the authors. Key to this will be auditing and testing pricing algorithms, looking for things like retaliatory pricing, price matching, and aggressive responses to price drops but not price rises.
Once collusive pricing rules are uncovered, computer scientists need to come up with ways to constrain algorithms from adopting them without sacrificing their clear efficiency benefits. It could also be helpful to make preventing this kind of collusive behavior the responsibility of the companies deploying them, with stiff penalties for those who don’t keep their algorithms in check.
One problem, though, is that algorithms may evolve strategies that humans would never think of, which could make spotting this behavior tricky. Imbuing courts with the technical knowledge and capacity to investigate this kind of evidence will also prove difficult, but getting to grips with these problems is an even more pressing challenge than it might seem at first.
While anti-competitive pricing algorithms could wreak havoc, there are plenty of other arenas where collusive AI could have even more insidious effects, from military applications to healthcare and insurance. Developing the capacity to predict and prevent AI scheming against us will likely be crucial going forward.
Image Credit: Pexels from Pixabay Continue reading
#437751 Startup and Academics Find Path to ...
Engineers have been chasing a form of AI that could drastically lower the energy required to do typical AI things like recognize words and images. This analog form of machine learning does one of the key mathematical operations of neural networks using the physics of a circuit instead of digital logic. But one of the main things limiting this approach is that deep learning’s training algorithm, back propagation, has to be done by GPUs or other separate digital systems.
Now University of Montreal AI expert Yoshua Bengio, his student Benjamin Scellier, and colleagues at startup Rain Neuromorphics have come up with way for analog AIs to train themselves. That method, called equilibrium propagation, could lead to continuously learning, low-power analog systems of a far greater computational ability than most in the industry now consider possible, according to Rain CTO Jack Kendall.
Analog circuits could save power in neural networks in part because they can efficiently perform a key calculation, called multiply and accumulate. That calculation multiplies values from inputs according to various weights, and then it sums all those values up. Two fundamental laws of electrical engineering can basically do that, too. Ohm’s Law multiplies voltage and conductance to give current, and Kirchoff’s Current Law sums the currents entering a point. By storing a neural network’s weights in resistive memory devices, such as memristors, multiply-and-accumulate can happen completely in analog, potentially reducing power consumption by orders of magnitude.
The reason analog AI systems can’t train themselves today has a lot to do with the variability of their components. Just like real neurons, those in analog neural networks don’t all behave exactly alike. To do back propagation with analog components, you must build two separate circuit pathways. One going forward to come up with an answer (called inferencing), the other going backward to do the learning so that the answer becomes more accurate. But because of the variability of analog components, the pathways don't match up.
“You end up accumulating error as you go backwards through the network,” says Bengio. To compensate, a network would need lots of power-hungry analog-to-digital and digital-to-analog circuits, defeating the point of going analog.
Equilibrium propagation allows learning and inferencing to happen on the same network, partly by adjusting the behavior of the network as a whole. “What [equilibrium propagation] allows us to do is to say how we should modify each of these devices so that the overall circuit performs the right thing,” he says. “We turn the physical computation that is happening in the analog devices directly to our advantage.”
Right now, equilibrium propagation is only working in simulation. But Rain plans to have a hardware proof-of-principle in late 2021, according to CEO and cofounder Gordon Wilson. “We are really trying to fundamentally reimagine the hardware computational substrate for artificial intelligence, find the right clues from the brain, and use those to inform the design of this,” he says. The result could be what they call end-to-end analog AI systems that capable of running sophisticated robots or even playing a role in data centers. Both of those are currently considered beyond the capabilities of analog AI, which is now focused only on adding inferencing abilities to sensors and other low-power “edge” devices, while leaving the learning to GPUs. Continue reading