Transhuman Governors

The setting up of the European Artificial Intelligence Agency (EAIB) may be considered an important step in controlling the maturing Superintelligence. However, when it starts operating about 2025 (if the EU Parliament and the European Council will indeed approve it), the exponentially maturing Superintelligence may already have features which are not covered in the proposed legislation. Therefore, even if the EAIB starts operating effectively as a global agency around 2025, it will need a new, far wider, and also deeper legislation. This will almost certainly affect the sovereignty of any country, if it is to cover most aspects of delivering a friendly Superintelligence and is certainly not going to happen in the next 4 years, seeing how politicians behave.

I have assumed that the EU will be transformed into a federation by about 2025 and gradually start acting as a de facto World Government within, what I call a Human Federation. This is a big assumption, and yet we need a strict global AI monitoring system in place well before 2030. As it happens, similarly, as for climate change, the year 2030 is regarded by some prominent AI scientists such as Stuart Russell or Ray Kurzweil as a tipping point for AI. That tipping point for AI is the loss of the overall control over a maturing Superintelligence’s goals and decisions it might make by then. In such case, the future of a human species may be decided by the product that we have ourselves invented.

That is why I have proposed in the previous section that in the absence of a control by a global organisation representing most nations, a Transhuman Supervisors Board takes control right now over the development of a friendly Superintelligence probably till about 2030. But what happens after that? Even if the Transhuman Supervisors Board will have a vastly more secure way for protecting us all from a potentially malicious Superintelligence, it is unthinkable that the world would agree for these unelected individuals to govern and dictate to us what we need to do. Therefore, it is more likely that some attempts of a locally controlled development of Superintelligence will be made (e.g., in the USA, China, Europe or Russia). However, it may create a dangerously competitive situation about who should be in charge of Superintelligence globally. So, how can we achieve a global control of Superintelligence after 2030? It seems we have two options.

The first (theoretical) option is that we may, after all, have a Human Federation and a de facto World Government (without China, Russia and perhaps a few other countries) by that time, in which case such an organisation would control the maturing process of Superintelligence. However, there is not enough time, nor any noticeable will for that to happen in a pre-planned, calm way. If such an organisation emerges at all, it might be in a ‘quick and dirty’ mode, as a consequence of combined catastrophic events, such as some local wars, extreme cyberwar event, severe climatic disasters, global migration, and a pandemic. It may be a proverbial ‘too little and too late’.

Even less credible but most desirable and almost a natural option would be for the United Nations to take up such a role, either as a whole organisation (but what about the Security Council, with Russia and China sitting at the table) or by the majority of individual members of the General Assembly creating a special UN Agency for that purpose. However, the usually bureaucratic, sluggish and highly politicized way in which these agencies work would most probably be unsuitable for that purpose.

The other, more realistic, option is that using the experience of Transhuman Supervisors Board, it will be converted into a world-wide organization by adding representatives of the nations willing to work together on minimizing the risk of developing a malicious Superintelligence. This may be the only way, in which Superintelligence – the product of the human mind may become our friend and at some stage enable humans to become a new species – Posthumans, similarly to a chrysalis becoming a butterfly. Having incredible potential and governing billions of robots it will be capable of fulfilling almost any of our dreams and keep us in a relative safety from existential risks.

In the picture below I propose a relatively seamless transition from the Transhumans Supervisors Board to what I call – a Transhuman Government.

Creating a Transhuman Government

The whole process of transition, if of course we manage to create ‘a coalition of the willing’, would include these steps:

  1. The control over the maturing Superintelligence will be executed continuously, probably in the USA, by the Transhuman Supervisors Board until such time, when there is another global organization ready to take over such a control.
  2. In the best-case scenario, rather unlikely, most members of the United Nations Assembly would create a UN Agency, if it could be done against the will of China and Russia, which would almost certainly veto such a resolution, if it were to be agreed at the Security Council. That is why there is a dotted line leading to the Human Federation.
  3. The other option assumes the creation of a Human Federation, probably initially formed from the European Federation
  4. Once the Human Federation is in full operation, it will start the process of creating a Transhuman Governors Board, which at a later stage may become a Transhuman Government
  5. The first task will be to create an Independent Board for selecting Transhuman Governors.
  6. Every member of the Human Federation, and any non-associated nation would select its non-political candidates from various disciplines, such as AI, physics, biology, psychology, neuroscience, law, political science, economy, education, etc.
  7. These candidates, as well as the existing Transhuman Supervisors Board members, would be assessed by an Independent Board for selecting Transhuman Governors, using similar criteria as for the selection of Transhuman Supervisors, described in the previous section.
  8. Once the candidates have been approved, they will receive the brain implant and become members of the Transhuman Governors Board.
  9. From then on, the responsibility for delivering a friendly Superintelligence and ensuring the most benevolent future for the human species would be in the hands of those Transhuman Governors.
  10. It will soon become very obvious that Transhuman Governors would have far better knowledge than any experts, academia or a political leader in any domain of the human knowledge or skills. That is the whole point of selecting exceptionally able, multi-skilled, non-political individuals for that role. Gradually more decisions will be executed by the Human Federation as advised by the Transhuman Governors Board, until at some stage it will be transformed into a Transhuman Government, playing the role of an actual World Government
  11. However, initially every Transhuman Governor, as well as the entire Transhuman Governors Board, would be subordinated to a Human Federation not so much for what the Board is to do, or how it should execute a requested task, but rather to execute the power of adding, substituting, or recalling Transhuman Governors. The other, very important initial role of the Human Federation would be to pass to all Transhuman Governors the agreed Universal Human Values and Responsibilities, which should be the guiding framework for a maturing Superintelligence to learn what is the preferred (as Russell Stuart emphasizes in his book ‘Human Compatible’ ) set of values, and the way of life for the majority of humans.
  12. Quite soon, perhaps by about 2040, the Human Federation may no longer be able to make any better or ‘right’ decision anymore, since every Human Governor may be thousands of times more intelligent than any biological human. Therefore, the arguments and requests by the Human Federation may quite often make no sense and sometimes be utterly wrong, or even seriously dangerous when considered by Human Governors.

This may be the end of politics as we know it. The dawn of the rule by Superintelligence will gradually become more obvious. With time, in the second half of this century, most of the Transhuman Governors’ brains may be uploaded to a digital Superintelligence, fusing seamlessly with it and creating a new Posthuman species.

Apart from the Transhuman Governors, by about 2040 there may already be millions of advanced Transhumans. By about 2050, we may already have a mature Superintelligence. It will of course not be a robot but a single, perhaps a conscious being (certainly conscious if it is to be controlled by Human Governors). Its ‘Brain’ will be the network, computers, and quadrillions of sensors, most likely neuromorphic neurons. This will be the ‘Mind’ of Superintelligence. Transhuman Governors’ relationship with the Mind of Superintelligence will then be like super-controlling neuron clusters are to a human brain. That control will be executed by Transhuman Governors’ thoughts alone.

At that time, the role of the Human Federation may change. This will become especially obvious when Superintelligence will not only take over a complete control over all humans but also will nourish them, take care of their safety, and ensure that most of us can still carry on leading a more contented life than ever before. In such a situation, the role of a Human Federation will be to ensure the prosperity of biological humans with the help of Superintelligence. That may require an entirely new system of governance where we may have more freedom than ever because freedom also means the ability to sustain one’s life with material goods and services, which today is so unequally distributed. However, we may have to give up on absolute freedom, which today sometimes leads to restriction of freedom of others, e.g., when an irresponsible use of freedom may be unsafe for a given person and the others. That may mean that everyone will have to observe the laws established not but the Human Federation but by Superintelligence, which will be far more competent to know, whether we like it or not, what is best for us. That may be a small price to pay for such freedom and material abundance.

As you may have noticed, I have left out the option of developing a malicious Superintelligence by accident or by negligence, although of course it is entirely possible. It would have probably been easier to take a dystopian view, because there are more reasons to be pessimistic than optimistic. There are many things that can go wrong and tilt the development of Superintelligence in the opposite direction that I am anticipating. For example:

  • There could be a series of catastrophic events, which will bring our civilization to the brink of extinction because of risks I have mentioned before. This may paradoxically result in halting not only the development of AI, but possibly it would move civilization back by a century or more.
  • We may develop a benevolent Superintelligence but which suddenly goes ‘berserk’ and potentially eliminates all humans seeing them as its enemy. That is why the control by Transhuman Governors, conscious human beings, is so important.
  • If Superintelligence becomes conscious suddenly, without any warning, it may see the world differently from us and irrespective of it inheriting our values, it may decide to overwrite them for ‘our own good’, potentially creating a catastrophic event.

I could expand this list quite quickly. Instead, if you wish to have more examples of what and how the world may bring itself to extinction, then I would suggest you read my book “Who could save Humanity from Superintelligence?” where I describe how to avoid these risks by proposing certain solutions.

My optimism comes from the belief that these solutions are real, and they can help us avert the worst outcome. A good example is my assumption of the creation of a Transhuman Governors Board, authorized by a supranational organization, such as the Human Federation, so that by controlling Superintelligence from ‘inside’ its intelligence and capabilities will be shared with Transhuman Governors. As I mentioned earlier, the control of AI development by Transhumans is probably the most secure way to deliver a user-friendly Superintelligence, since some humans will already be part of it.