Is March 21, 2020 still a good date for a global referendum to: 1) affirm that the collective will of humanity is the ultimate expression of political power on planet Earth, 2) give the principles of the Universal Declaration of Human Rights and the Earth Charter the force of supreme law for all of humanity, and 3) approve a plan for immediate worldwide demilitarization?
It seems to me that the answer is Yes. Part of the reason is that the number of people doing research on machine learning (ML), from countries all around the world, is growing at an exponential rate as is the speed of the hardware. A machine learning model can be trained in hours on the hardware of today when it might have taken months to train on the hardware of four or five years ago. State of the art ML hardware for mobile devices has been announced by Arm that will appear in products by the end of 2019.
The comparison of machine intelligence in early 2020 with what we experience today will mainly be through our digital assistants, also known as avatars (Grandi, 2018). Computer speech recognition and language translation are already pretty good. It is in the domain of question answering that the largest increases in machine intelligence are yet to be seen.
Well, that’s all pretty mechanical. Is there something else, human consciousness for example, that needs to be taken into consideration as we work toward the March 21, 2020 revolution?
What is the state of human consciousness? To what extent do we realize that we are connected, and at the most basic level, one with each other and with the rest of the universe, infinite and eternal?
Maybe that is too much to ask for. It might be sufficient to achieve some awareness that the whole is greater than the sum of the parts, and that this applies to humanity, as well as to humanity in relation to the rest of the biosphere. Or, perhaps what is required is for enough people to see that liberating ourselves from this corrupt, hierarchical structure of power is to our personal benefit, and this means uniting to make it happen.
There have been a number of works of fiction, Ender’s Game by Orson Scott Card, for example, in which humanity has been able to unite, at least temporarily, when confronted with an invading, extraterrestrial force. The buggers, large insect-appearing aliens in the story of Ender, arrived in the solar system, and began slaughtering large sectors of humanity. The surviving humans united to fight back.
The thought has been expressed, almost wistfully, that it would be good if we actually were confronted with some kind of extraterrestrial threat in order to bring about the unification of humanity as is so desperately needed. No such invasion has occurred.
Maybe the existential threat that could result in the unification of humanity is global warming. Maybe not. Even though the current concentration of atmospheric greenhouse gases is changing the global climate incredibly quickly, from a geological perspective, from the perspective of a human lifetime, the change is happening so slowly that some people are still in denial.*
Let’s go back to the idea of an existential threat posed by an extraterrestrial intelligence, except this time, let’s think about it in terms of an emerging terrestrial intelligence, machine intelligence. There are some people who do consider machine intelligence to be an existential threat, and maybe the worst one that is facing humanity. It may be a realistic fear.
It’s a matter of consciousness. Are we intelligent enough to turn it around, and look at it from another angle? Suppose that what machine intelligence poses to humanity is an existential opportunity.
There are videos at YouTube of researchers giving presentations at conferences on machine learning. One can tell that they are not motivated by a political agenda. They are working to expand the boundaries of what is possible in this new frontier. Some might harbor the hope that they will be remembered as another Turing, or Edison, or da Vinci. Who will contribute to the next leap in machine intelligence?
As noted above, it’s already pretty good at speech recognition and language translation, both of which are important for overcoming the barriers that divide humanity. Suppose that it improves dramatically in the next eighteen months at question answering. What questions would we want to ask?
If it could process the input from masses of people, I would like to know if it could discern a plan for immediate worldwide demilitarization behind which humanity could unite. Who would do the inspections for compliance? What alternative employment could be offered to the people who are currently bearing arms and working at arms manufacturing? How, specifically, would the trillions of dollars saved be spent? Where would we find the points of greatest unity among the billions of comments and opinions expressed?
According to some estimates, about four billion people are currently using the Internet, the vast majority of them with mobile phones. The revolution could be facilitated by some of these citizens interviewing those who do not yet have Internet access so their input could be included. Fortunately, there are multiple ways that the fewer than four billion people who are not already using the Internet can participate: demonstrations, strikes, occupy tactics, and voting with paper ballots. I expect that the billions who do not yet have Internet access will want to participate. They are the ones who have the most to gain.
While machine learning researchers are mainly good-hearted about it, what is to prevent the power elite from subverting the emerging machine intelligence, and using it to impose some awful totalitarian state? One thing we can try to do is make sure that the algorithms and software are open source, and published on the Internet in human readable format. Another is to strive to create a world government that is based on the rule of principles instead of personalities.
The immediate reward of worldwide demilitarization would be huge. I think that humanity as a whole is smart enough to see it, and the sooner we can make it so, the better. We don't want the emerging machine intelligence to be subverted to some evil purpose by the existing structure of power. March 21, 2020 still sounds like a good date for a global referendum to me. It's when the impossible may become possible.
July 12, 2018
* The asterisk at the end of the paragraph about the threat of global warming is to note that the denial might mainly be the result of political opposition instead of a lack of evidence by now.
U. Grandi. Agent-Mediated Social Choice. arXiv:1806.07199v2 [cs.AI] 10 Jul 2018. URL https://arxiv.org/pdf/1806.07199v2.pdf