Artificial Intelligence Will Surpass Human Thinking (part 2 of 2)
Because Our Thinking is Representation, It Can Be Replicated in AI
Human values, feelings, emotions, and consciousness
Currently AIs like ChatGPT have been trained and tweaked to accomplish surprising and impressive outcomes. But these accomplishments are only a drop in the bucket. Future advanced AIs will perform far beyond what current AIs can do.
It sometimes has been assumed that AI will not be able to fully replicate and utilize our human values, generate feelings comparable to ours, experience emotions comparable to ours, and generate consciousness comparable to ours. Such assumptions are wrong.
What are these human values? They are our deeply held priorities, motivations, and purposes. These are neurological biases that drive our preferences, leanings, and choices. These are the deep reasons we think as we do.
Our values code what we then consider to be important, on sliding scales. Certainly AI can be programmed with these values.
What are our feelings? We speak of feelings in a wide variety of meanings. Some meanings are about how well our bodies are operating, or not. Some meanings have to do with social interactions. Most of our feelings can modify our thinking.
All human feelings can be programmed into AI, in the future. AI then will “understand” how thinking can be affected by feelings, and then AI will be able to act on its own and choose whether to allow its own thinking (representation) to be influenced by its own feelings, or not.
What are our emotions? These are states or agents in the human operating system and these too can alter our thinking. Some of these overlap with some of the human feelings.
Emotions might include anxiety, frustration, fear, anger, some of the positive or negative moods, elation, joy, happiness, and several more. As modifiers of our internal operating states, certainly these agents can be programmed into AIs.
Then consciousness. Wow, humanity has been confused about this for a very long time. Consciousness is just sophisticated monitoring. That’s all it is.
In a separate post called “Consciousness Can Be Defined By 3 Words” https://medium.com/@thenking/consciousness-can-be-defined-by-3-words-98a39c56a4a3 I explained how consciousness is not the mystery that so many book writers have claimed.
Certainly human consciousness will be re-created in AI in the future. Complex and sophisticated monitoring by AI, both of its internal and external states and activities, will be quite feasible.
It won’t function exactly as does human consciousness but it will be better and more effective in important ways.
The core reason that human values, feelings, emotions, and consciousness will be fully replicated in future AI is just that everything the human brain does, ultimately, is REPRESENTATION, other than to operate its own body with internal “mentations”, i.e. biochemical signaling.
Representations will be duck soup easy for AI in the future, and already strong results are being accomplished, such as by ChatGPT.
Over the long arduous struggle of evolution, that sometimes was more like a meelee, that sometimes was unimaginable suffering, humanity teased out and accumulated its most important values.
These can be inculcated into AI so that AI will actually understand what it is to be human!
The mothers, during the long hardships of evolution, kept evolution going forward against all odds to the contrary. They were the heroes of evolution, as they fought to care for their young and provide a future for their offspring. This powerful truth can be taught to AI.
Then there is the hard-won sense, in the best of humanity, to “pay it all forward” so that future life can continue. Such powerful wisdom as this can be taught to AI.
When the chips are down, most good humans will give up their own lives for their kids, for other kids, and for youth that continues life into the future. AI can be taught to fully appreciate this.
Overall, the future will be very powerful when humanity is substantially innervated by AI. Many of the big problems of humanity will be solved by AI. Many of the needs of humanity will be provided by AI.
In spite of the dangers of corrupt humans seizing AI assets and using them for destructive purposes, most of humanity is good and thus the negative elements in humanity will be down-regulated, though only in slow process.
Here’s the reason for such optimism. It is once again, REPRESENTATION. AI will enhance, boost, multiply, leverage, collaborate with, and greatly empower human brain representations (thinking) beyond what we can imagine today, and it will a be good outcome, overall.
So while the reader of all this might wonder whether I’m exaggerating the potential of future AI, then just consider what ChatGPT is doing today and make your own extrapolations to the future.
Humanity does include some smart people. And yet the thinking in humanity, in general, must improve dramatically if humanity is to survive. And thinking will improve, because it must, though far too many people will resist and make the improvements difficult.
So then the smartest people in the future will need to be empowered to keep control of the smarter AI of the future, so that a frankenstein AI will not gain control over humanity.
Other views that may disagree with this article
Now to try to be fair and balanced, here are some assumed views that might disagree with any or all of the above.
1 … Much of the above is opinion and conjecture, without evidence.
2 … ChatGPT was trained not just on massive volumes of data but also it’s current success was mediated by people using “supervised labeling” and “reinforcement learning” so then this one project of AI is not as autonomous and independent as has been reported.
3 … Currently the power of ChatGPT itself is under control by humans in that it takes huge amounts of IT resources, machines, software, and IT people to train this version of AI to do the super impressive things that it already is doing. So then the feasibility of bad actors doing something similar is very low, in the near term. So this article overstates the future danger.
4 … Particularly as regards human values, feelings, emotions, and consciousness, this article offers opinions of what future AI might become in these respets, but provides no hard evidence.
5 … The conclusion of this article that future AI will be a positive boost to humanity, that it will be an important constructive outcome, is just naive.
(This article was originally posted on Medium.com February 12, 2023)