Vivek Wadhwa and Mauritz Kop
Today’s artificial intelligence is as self-aware as a paper clip. Despite the hype—such as a Google engineer’s bizarre claim that his company’s AI system had “come to life” and Tesla CEO Elon Musk’s tweet predicting that computers will have human intelligence by 2029—the technology still fails at simple everyday tasks. That includes driving vehicles, especially when confronted by unexpected circumstances that require even the tiniest shred of human intuition or thinking.
The sensationalism surrounding AI is not surprising, considering that Musk himself had warned that the technology could become humanity’s “biggest existential threat” if governments don’t regulate it. But whether or not computers ever attain human-like intelligence, the world has already summoned a different, equally destructive AI demon: Precisely because today’s AI is little more than a brute, unintelligent system for automating decisions using algorithms and other technologies that crunch superhuman amounts of data, its widespread use by governments and companies to surveil public spaces, monitor social media, create deepfakes, and unleash autonomous lethal weapons has become dangerous to humanity.
Compounding the danger is the lack of any AI regulation. Instead, unaccountable technology conglomerates, such as Google and Meta, have assumed the roles of judge and jury in all things AI. They are silencing dissenting voices, including their own engineers who warn of the dangers.
The world’s failure to rein in the demon of AI—or rather, the crude technologies masquerading as such—should serve to be a profound warning. There is an even more powerful emerging technology with the potential to wreak havoc, especially if it is combined with AI: quantum computing. We urgently need to understand this technology’s potential impact, regulate it, and prevent it from getting into the wrong hands before it is too late. The world must not repeat the mistakes it made by refusing to regulate AI.
Although still in its infancy, quantum computing operates on a very different basis from today’s semiconductor-based computers. If the various projects being pursued around the world succeed, these machines will be immensely powerful, performing tasks in seconds that would take conventional computers millions of years to conduct.
Because of the technology’s immense power and revolutionary applications, quantum computing projects are likely part of defense and other government research already.
Semiconductors represent information as a series of 1s and 0s—that’s why we call it digital technology. Quantum computers, on the other hand, use a unit of computing called a qubit. A qubit can hold values of 1 and 0 simultaneously by incorporating a counterintuitive property in quantum physics called superposition. (If you find this confusing, you’re in good company—it can be hard to grasp even for experienced engineers.) Thus, two qubits could represent the sequences 1-0, 1-1, 0-1, and 0-0, all in parallel and all at the same instant. That allows a vast increase in computing power, which grows exponentially with each additional qubit.
If quantum physics leaves the experimental stage and makes it into everyday applications, it will find many uses and change many aspects of life. With their power to quickly crunch immense amounts of data that would overwhelm any of today’s systems, quantum computers could potentially enable better weather forecasting, financial analysis, logistics planning, space research, and drug discovery. Some actors will very likely use them for nefarious purposes, compromising bank records, private communications, and passwords on every digital computer in the world. Today’s cryptography encodes data in large combinations of numbers that are impossible to crack within a reasonable time using classic digital technology. But quantum computers—taking advantage of quantum mechanical phenomena, such as superposition, entanglement, and uncertainty—may potentially be able to try out combinations so rapidly that they could crack encryptions by brute force almost instantaneously.
To be clear, quantum computing is still in an embryonic stage—though where, exactly, we can only guess. Because of the technology’s immense potential power and revolutionary applications, quantum computing projects are likely part of defense and other government research already. This kind of research is shrouded in secrecy, and there are a lot of claims and speculation about milestones being reached. China, France, Russia, Germany, the Netherlands, Britain, Canada, and India are known to be pursuing projects. In the United States, contenders include IBM, Google, Intel, and Microsoft as well as various start-ups, defense contractors, and universities.
Despite the lack of publicity, there have been credible demonstrations of some basic applications, including quantum sensors able to detect and measure electromagnetic signals. One such sensor was used to precisely measure Earth’s magnetic field from the International Space Station.
Dutch researchers teleported quantum information in another experiment across a rudimentary quantum communication network. Instead of using conventional optical fibers, the scientists used three small quantum processors to instantly transfer quantum bits from a sender to a receiver. These experiments haven’t shown practical applications yet, but they could lay the groundwork for a future quantum internet, where quantum data can be securely transported across a network of quantum computers faster than the speed of light. So far, that’s only been possible in the realm of science fiction.
The Biden administration considers the risk of losing the quantum computing race imminent and dire enough that it issued two presidential directives in May: one to place the National Quantum Initiative advisory committee directly under the authority of the White House and another to direct government agencies to ensure U.S. leadership in quantum computing while mitigating the potential security risks quantum computing poses to cryptographic systems.
Experiments are also working to combine quantum computing with AI to transcend traditional computers’ limits. Today, large machine-learning models take months to train on digital computers because of the vast number of calculations that must be performed—OpenAI’s GPT-3, for example, has 175 billion parameters. When these models grow into the trillions of parameters—a requirement for today’s dumb AI to become smart—they will take even longer to train. Quantum computers could greatly accelerate this process while also using less energy and space. In March 2020, Google launched TensorFlow Quantum, one of the first quantum-AI hybrid platforms that takes the search for patterns and anomalies in huge amounts of data to the next level. Combined with quantum computing, AI could, in theory, lead to even more revolutionary outcomes than the AI sentience critics have warned about.
Given the potential scope and capabilities of quantum technology, it is absolutely crucial not to repeat the mistakes made with AI—where regulatory failure has given the world algorithmic bias that hypercharges human prejudices, social media that favors conspiracy theories, and attacks on the institutions of democracy fueled by AI-generated fake news and social media posts. The dangers lie in the machine’s ability to make decisions autonomously, with flaws in the computer code resulting in unanticipated, often detrimental, outcomes. In 2021, the quantum community issued a call for action to urgently address these concerns. In addition, critical public and private intellectual property on quantum-enabling technologies must be protected from theft and abuse by the United States’ adversaries.
There are national defense issues involved as well. In security technology circles, the holy grail is what’s called a cryptanalytically relevant quantum computer—a system capable of breaking much of the public-key cryptography that digital systems around the world use, which would enable blockchain cracking, for example. That’s a very dangerous capability to have in the hands of an adversarial regime.
Experts warn that China appears to have a lead in various areas of quantum technology, such as quantum networks and quantum processors. Two of the world’s most powerful quantum computers were built in China, and as far back as 2017, scientists at the University of Science and Technology of China in Hefei built the world’s first quantum communication network using advanced satellites. To be sure, these publicly disclosed projects are scientific machines to prove the concept, with relatively little bearing on the future viability of quantum computing. However, knowing that all governments are pursuing the technology simply to prevent an adversary from being first, these Chinese successes could well indicate an advantage over the United States and the rest of the West.
Beyond accelerating research, targeted controls on developers, users, and exports should therefore be implemented without delay. Patents, trade secrets, and related intellectual property rights should be tightly secured—a return to the kind of technology control that was a major element of security policy during the Cold War. The revolutionary potential of quantum computing raises the risks associated with intellectual property theft by China and other countries to a new level.
Finally, to avoid the ethical problems that went horribly wrong with AI and machine learning, democratic nations need to institute controls that correspond to the power of the technology and respect democratic values, human rights, and fundamental freedoms. Governments must urgently consider regulations, standards, and responsible uses—and learn from how countries handled or mishandled other revolutionary technologies, including AI, nanotechnology, biotechnology, semiconductors, and nuclear fission. The United States and other democratic nations must not make the same mistake they made with AI—and prepare for tomorrow’s quantum era today.
No comments:
Post a Comment