Tag Archives: Artificial intelligence
#431337 Church that Worships AI God May Be the ...
A former Google and Uber engineer wants to establish Way of the Future, a religious group dedicated to a "godhead based on artificial intelligence." Continue reading
#431300 Benefits and Risks of Artificial ...
Introduction Robots have been part of the manufacturing industry for longer than most people are aware, but the advent of linear actuators has created a world in which they can be used in more industries and workplaces throughout the world. People aren’t stopping with robots in more minor areas – they are thinking of bringing … Continue reading
#431243 Does Our Survival Depend on Relentless ...
Malthus had a fever dream in the 1790s. While the world was marveling in the first manifestations of modern science and technology and the industrial revolution that was just beginning, he was concerned. He saw the exponential growth in the human population as a terrible problem for the species—an existential threat. He was afraid the human population would overshoot the availability of resources, and then things would really hit the fan.
“Famine seems to be the last, the most dreadful resource of nature. The power of population is so superior to the power of the earth to produce subsistence for man, that premature death must in some shape or other visit the human race. The vices of mankind are active and able ministers of depopulation.”
So Malthus wrote in his famous text, an essay on the principles of population.
But Malthus was wrong. Not just in his proposed solution, which was to stop giving aid and food to the poor so that they wouldn’t explode in population. His prediction was also wrong: there was no great, overwhelming famine that caused the population to stay at the levels of the 1790s. Instead, the world population—with a few dips—has continued to grow exponentially ever since. And it’s still growing.
There have concurrently been developments in agriculture and medicine and, in the 20th century, the Green Revolution, in which Norman Borlaug ensured that countries adopted high-yield varieties of crops—the first precursors to modern ideas of genetically engineering food to produce better crops and more growth. The world was able to produce an astonishing amount of food—enough, in the modern era, for ten billion people. It is only a grave injustice in the way that food is distributed that means 12 percent of the world goes hungry, and we still have starvation. But, aside from that, we were saved by the majesty of another kind of exponential growth; the population grew, but the ability to produce food grew faster.
In so much of the world around us today, there’s the same old story. Take exploitation of fossil fuels: here, there is another exponential race. The exponential growth of our ability to mine coal, extract natural gas, refine oil from ever more complex hydrocarbons: this is pitted against our growing appetite. The stock market is built on exponential growth; you cannot provide compound interest unless the economy grows by a certain percentage a year.
“This relentless and ruthless expectation—that technology will continue to improve in ways we can’t foresee—is not just baked into share prices, but into the very survival of our species.”
When the economy fails to grow exponentially, it’s considered a crisis: a financial catastrophe. This expectation penetrates down to individual investors. In the cryptocurrency markets—hardly immune from bubbles, the bull-and-bear cycle of economics—the traders’ saying is “Buy the hype, sell the news.” Before an announcement is made, the expectation of growth, of a boost—the psychological shift—is almost invariably worth more than whatever the major announcement turns out to be. The idea of growth is baked into the share price, to the extent that even good news can often cause the price to dip when it’s delivered.
In the same way, this relentless and ruthless expectation—that technology will continue to improve in ways we can’t foresee—is not just baked into share prices, but into the very survival of our species. A third of Earth’s soil has been acutely degraded due to agriculture; we are looming on the brink of a topsoil crisis. In less relentless times, we may have tried to solve the problem by letting the fields lie fallow for a few years. But that’s no longer an option: if we do so, people will starve. Instead, we look to a second Green Revolution—genetically modified crops, or hydroponics—to save us.
Climate change is considered by many to be an existential threat. The Intergovernmental Panel on Climate Change has already put their faith in the exponential growth of technology. Many of the scenarios where they can successfully imagine the human race dealing with the climate crisis involve the development and widespread deployment of carbon capture and storage technology. Our hope for the future already has built-in expectations of exponential growth in our technology in this field. Alongside this, to reduce carbon emissions to zero on the timescales we need to, we will surely require new technologies in renewable energy, energy efficiency, and electrification of the transport system.
Without exponential growth in technology continuing, then, we are doomed. Humanity finds itself on a treadmill that’s rapidly accelerating, with the risk of plunging into the abyss if we can’t keep up the pace. Yet this very acceleration could also pose an existential threat. As our global system becomes more interconnected and complex, chaos theory takes over: the economics of a town in Macedonia can influence a US presidential election; critical infrastructure can be brought down by cybercriminals.
New threats, such as biotechnology, nanotechnology, or a generalized artificial intelligence, could put incredible power—power over the entire species—into the hands of a small number of people. We are faced with a paradox: the continued existence of our system depends on the exponential growth of our capacities outpacing the exponential growth of our needs and desires. Yet this very growth will create threats that are unimaginably larger than any humans have faced before in history.
“It is necessary that we understand the consequences and prospects for exponential growth: that we understand the nature of the race that we’re in.”
Neo-Luddites may find satisfaction in rejecting the ill-effects of technology, but they will still live in a society where technology is the lifeblood that keeps the whole system pumping. Now, more than ever, it is necessary that we understand the consequences and prospects for exponential growth: that we understand the nature of the race that we’re in.
If we decide that limitless exponential growth on a finite planet is unsustainable, we need to plan for the transition to a new way of living before our ability to accelerate runs out. If we require new technologies or fields of study to enable this growth to continue, we must focus our efforts on these before anything else. If we want to survive the 21st century without major catastrophe, we don’t have a choice but to understand it. Almost by default, we’re all accelerationists now.
Image Credit: focal point / Shutterstock.com Continue reading
#431238 AI Is Easy to Fool—Why That Needs to ...
Con artistry is one of the world’s oldest and most innovative professions, and it may soon have a new target. Research suggests artificial intelligence may be uniquely susceptible to tricksters, and as its influence in the modern world grows, attacks against it are likely to become more common.
The root of the problem lies in the fact that artificial intelligence algorithms learn about the world in very different ways than people do, and so slight tweaks to the data fed into these algorithms can throw them off completely while remaining imperceptible to humans.
Much of the research into this area has been conducted on image recognition systems, in particular those relying on deep learning neural networks. These systems are trained by showing them thousands of examples of images of a particular object until they can extract common features that allow them to accurately spot the object in new images.
But the features they extract are not necessarily the same high-level features a human would be looking for, like the word STOP on a sign or a tail on a dog. These systems analyze images at the individual pixel level to detect patterns shared between examples. These patterns can be obscure combinations of pixel values, in small pockets or spread across the image, that would be impossible to discern for a human, but highly accurate at predicting a particular object.
“An attacker can trick the object recognition algorithm into seeing something that isn’t there, without these alterations being obvious to a human.”
What this means is that by identifying these patterns and overlaying them over a different image, an attacker can trick the object recognition algorithm into seeing something that isn’t there, without these alterations being obvious to a human. This kind of manipulation is known as an “adversarial attack.”
Early attempts to trick image recognition systems this way required access to the algorithm’s inner workings to decipher these patterns. But in 2016 researchers demonstrated a “black box” attack that enabled them to trick such a system without knowing its inner workings.
By feeding the system doctored images and seeing how it classified them, they were able to work out what it was focusing on and therefore generate images they knew would fool it. Importantly, the doctored images were not obviously different to human eyes.
These approaches were tested by feeding doctored image data directly into the algorithm, but more recently, similar approaches have been applied in the real world. Last year it was shown that printouts of doctored images that were then photographed on a smartphone successfully tricked an image classification system.
Another group showed that wearing specially designed, psychedelically-colored spectacles could trick a facial recognition system into thinking people were celebrities. In August scientists showed that adding stickers to stop signs in particular configurations could cause a neural net designed to spot them to misclassify the signs.
These last two examples highlight some of the potential nefarious applications for this technology. Getting a self-driving car to miss a stop sign could cause an accident, either for insurance fraud or to do someone harm. If facial recognition becomes increasingly popular for biometric security applications, being able to pose as someone else could be very useful to a con artist.
Unsurprisingly, there are already efforts to counteract the threat of adversarial attacks. In particular, it has been shown that deep neural networks can be trained to detect adversarial images. One study from the Bosch Center for AI demonstrated such a detector, an adversarial attack that fools the detector, and a training regime for the detector that nullifies the attack, hinting at the kind of arms race we are likely to see in the future.
While image recognition systems provide an easy-to-visualize demonstration, they’re not the only machine learning systems at risk. The techniques used to perturb pixel data can be applied to other kinds of data too.
“Bypassing cybersecurity defenses is one of the more worrying and probable near-term applications for this approach.”
Chinese researchers showed that adding specific words to a sentence or misspelling a word can completely throw off machine learning systems designed to analyze what a passage of text is about. Another group demonstrated that garbled sounds played over speakers could make a smartphone running the Google Now voice command system visit a particular web address, which could be used to download malware.
This last example points toward one of the more worrying and probable near-term applications for this approach: bypassing cybersecurity defenses. The industry is increasingly using machine learning and data analytics to identify malware and detect intrusions, but these systems are also highly susceptible to trickery.
At this summer’s DEF CON hacking convention, a security firm demonstrated they could bypass anti-malware AI using a similar approach to the earlier black box attack on the image classifier, but super-powered with an AI of their own.
Their system fed malicious code to the antivirus software and then noted the score it was given. It then used genetic algorithms to iteratively tweak the code until it was able to bypass the defenses while maintaining its function.
All the approaches noted so far are focused on tricking pre-trained machine learning systems, but another approach of major concern to the cybersecurity industry is that of “data poisoning.” This is the idea that introducing false data into a machine learning system’s training set will cause it to start misclassifying things.
This could be particularly challenging for things like anti-malware systems that are constantly being updated to take into account new viruses. A related approach bombards systems with data designed to generate false positives so the defenders recalibrate their systems in a way that then allows the attackers to sneak in.
How likely it is that these approaches will be used in the wild will depend on the potential reward and the sophistication of the attackers. Most of the techniques described above require high levels of domain expertise, but it’s becoming ever easier to access training materials and tools for machine learning.
Simpler versions of machine learning have been at the heart of email spam filters for years, and spammers have developed a host of innovative workarounds to circumvent them. As machine learning and AI increasingly embed themselves in our lives, the rewards for learning how to trick them will likely outweigh the costs.
Image Credit: Nejron Photo / Shutterstock.com Continue reading