Solar Carport System,Solar Carport Structural,Carport Solar Parking,Custom Car Parking Solar Carport Structural Hebei Jinbiao Construction Materials Tech Corp., Ltd. , https://www.pvcarportsystem.com
Robot learns to lie? People's understanding of AI is too exaggerated
[NetEase Smart News, August 19] Today, artificial intelligence (AI) has become ubiquitous. Products like Amazon's Alexa, Google's Google Assistant, Apple's Siri, and Microsoft's Cortana are all examples of AI in action. However, speech recognition isn't exactly groundbreaking anymore. Software like Dragon has been steadily improving for over 20 years.
Recently, there has been much debate surrounding AI, with claims of new breakthroughs like "deep learning" potentially replacing over 2 million Australian workers by 2030. Similar predictions have been made before.
When I led the team developing shearing robots in the 1970s, I had the privilege of discussing AI with philosopher Julius 燢ovesi. His insights were profound. He argued that robots are fundamentally similar to ordinary toilet tanks, merely simple automated devices.
He once said, "If you can show me a robot that manipulates your behavior through intentional deception, then I'll admit it has artificial intelligence!" Of course, this is precisely what we strive for in our shearing robots.
To better understand the future, it helps to view AI as another method of programming digital computers. This is the current reality. For decades, we've been learning to coexist with computers, growing increasingly reliant on them while they become more user-friendly. Smartphones are a prime example.
Our work has changed and will continue to change. Yet, there are exaggerated claims about AI. In the 1960s, natural language machine translation was predicted to be "just two to three years away." We're still far from achieving this. In the late 1970s and early 1980s, many believed that by the mid-1990s, 95% of factory jobs would vanish. Again, we're still waiting for this to happen. The "dot-com bubble" of 2001 sparked a new wave of AI hype, but as always, these arguments faded and disappointment followed.
Self-driving cars will eventually grace our streets, thanks to decades of advancements in sensor technology, computer hardware, and software engineering. They'll begin cautiously but will improve over time. You might call this AI, but it won't fundamentally alter everything.
The real casualties here are our appreciation of human intelligence. AI merely mimics skilled gameplay, mathematical proofs, or legal and medical interpretations—activities associated with wise individuals. Consider simpler tasks like telling jokes or assessing safety. Cognitive scientists are still unraveling how we perform these.
Even animals challenge us. MIT researchers developed an artificial mine-detection device, which detected TNT everywhere during tests. Meanwhile, trained dogs quickly located mines. Imagine a crowded room where someone hides a cigarette. While humans struggle to identify the source, a trained dog finds it instantly.
Some speculate that quantum computers could revolutionize AI, but for now, their experiments resemble Alan Turing's relay adjustments in the 1920s. We're still far from knowing if these machines can deceive. Moreover, it's worth questioning whether today's AI boom, driven by companies like Google and Facebook, is merely investor bait. Perhaps it's just another form of collective wishful thinking. [Source: afr.com, Compiled by James Trevelyan, Translated by Netease External Compilation Platform, Edited by Wayne]
At times, it feels like AI is both a promise and a mirage. While the technology continues to evolve, it remains essential to separate hype from reality. As we navigate this ever-changing landscape, let's remember that true innovation often lies in understanding—and respecting—the limits of what we create.