The aggrandizing of artificial intelligence appears to have a firm hold these days. Maybe you didn’t read Issac Azimov’s “I, Robot” series of books. I thoroughly enjoyed them as a kid. If you missed out, central to the stories were key rules that boil down to intelligent machines always putting human safety and rights first. It was plain to me this common sense rule wasn’t fiction – and that it applied to humans just as much to robots. But stories aside, just what is it about AI that prompts so much huzzah? Is it the signal that something’s been over-looked?
Is Machine Intelligence Really the Point?
With so many involved in AI development and the countless hours put into anticipating and executing purpose, design, and programming, it seems more likely to me by the day that comparing machine and human intelligence is inappropriate. The machines wouldn’t exist without the prerequisite contribution of human cognition. And programmers, just to consider one group of key creators, know better than most that garbage in equals garbage out. Successful AI is reliant on unerring human input.
Speed and Complexity Come Second
All computing machines have this in common: programming that follows exact rules. Errors all mean the same thing – the rules weren’t followed. And if they weren’t followed, the same action is applied in every case. Someone figures out which rules need attention and corrects things. It’s no different than doing our arithmetic. A faulty answer isn’t the point. It’s simply the evidence that we either didn’t understand – or didn’t follow – the rules. The task is then to learn them better or make sure we hold to them in future.
Whether we’re balancing the cheque book or trying to understand reality, it doesn’t matter how fast the calculating or how persuasive or fancy the theories. If the mark is missed and it isn’t noticed, chaos enters the system.
Your tablet and the latest and greatest super-computer use the same core instructions for essential functions. Without these basic directions, neither machine will get the job done properly, no matter how simple the task.
And this is where the machines clearly rate far higher than people. Are the ways we’re taught about the rules of reasoning any more consistent than the weather? What could we gain by applying the dedication and quality control heaped on technological development, to the advancement of all areas of human reasoning?
Besides not having complex biology to maintain, the machines reap the benefits of uniform, correctly thought out rules, which when followed, can’t help but cause them to excel at coming to correct conclusions.
Of Robots and People
The machines are faster at many tasks and can be programmed to deal with complexity that, if we had to do it ourselves, would eat up so much time there’d be none left for generating new ideas – even if we didn’t make any boo boos. Since machines are unlikely to ever match human capacity for creating, their function simply isn’t enough to replace us, no matter how advanced the calculating they eventually tap out at.
Rather than worry about an AI “take-over” it might be a good time to widen the view-point.
It’s noted that we tend to fear, demonize, or in opposite fashion, over-esteem – even worship – what we don’t yet understand. The latter detour might remind you of another sci-fi writer – Arthur C Clarke. To paraphrase his observation:
Technology sufficiently advanced appears as magic.
Looking at the machines capacity as magic has it’s difficulties.
When put aside, reality comes into focus. We can be glad that the machines take so much of the tedious and nit-picky load off. Just how valuable the rules are when learning to understand – not just grasping the outcomes we’re aiming for, but really getting a handle on the most sensible ways to implement them – becomes obvious. And another thing that comes quickly to mind is that the machine kind of accuracy is quite possible in every area of human endeavor.
Successful AI is Due to Unbiased, Effective Human Thinking
The development of intelligent machines serves as proof of concept. Just look at what has been accomplished by applying consistent, well understood rules! Any system can use these methods to reliably generate coherence.
But what about our “non-technological” rules? Rule breaking continually adds chaos to our systems. And since many human rules and laws are inconsistent, following the ones that haven’t been thought out properly also increase chaos.
Yet development of AI proceeds apace despite the messy social realm rules. How can this be? No, it isn’t magic. Developers just put the core rules first, and in doing so, the AI version of common sense is automatically maintained.
It’s Not Sci-Fi
Clarke’s idea isn’t too shabby, but Azimov hit the nail on the head with his robot laws. These rules fit right in with human core laws already in place internationally. All that’s left to do is put these rules in first place, no matter what area of reasoning or expertise we’re operating in. Human safety and rights laws – square one – must come ahead of all other rules and practices. This action makes it so much easier to de-bias. As errors are removed, the way to improvement opens wide. More and more incoherent rules can be straightened out or eliminated.
The initial gains we can look forward to? Even if our thinking doesn’t speed up – although I suspect it will – there will be fewer, but much more effective, ideas generated. Decision-making will become rock-solid. As these outcomes reach the mainstream, more and more time for error-free thinking and deciding will become available. Chaos will be increasingly constrained, allowing even more time for quality thought.
The smart machines aren’t our competitors. They are intended to extend our cognition and capability, not replace it. And, although quite a few hastily assume that rules are the bane of creativity, used properly, they actually are the foundation of it. Intelligent machines, and the careful procedures that make them possible, are both the blueprints for, and the finest, most helpful tools. And they’re increasingly being made ready to be put to their highest and best use.
Transfer this educational wealth. The best methods for using rules in common are already tested out in the realm of technology. Make this kind of standardization accessible as the global learning-path with human rights as the core. All hands on deck.