Based on the Humanity’s Mission, which I have proposed earlier (Avoid extinction by coexisting with Superintelligence and gradually evolve into a new species), we should have a Road Map for Humanity’s Evolution, with perhaps the following End Goal:
Teach and instil in AI the best human values and preferences until its mature form, Superintelligence, becomes a single entity, thousands of times more intelligent than humans and yet remaining our partner
Ideally, it should be implemented in this, most desired order, however realistically, the order and the scope of some goals is bound to change:
- Create Safe AI. Create a global AI-Governance legal framework, based most likely on the EU proposals, and implemented in a similar way to GDPR. This must be a starting point in the Road Map because it will create a de facto standard legal framework for controlling the development and maturing of AI. To be effectively implemented it needs:
- Global coverage. There should be a single Global AI Governance Agency, having control over any aspect of AI development, similar to IEEE or International Atomic Energy Agency (IAEA). Such control should extend over all states and individuals. Stuart Russell in his latest book ‘Human Compatible’, believes this should happen latest by 2030. This is also the date I put in my previous book, where I call the decade 2020-2030 – an age of Immature Superintelligence.
- Continuity. The control over AI development should begin right now and be continuously maintained until a fully mature Superintelligence emerges – quite probably by about 2050. Once we reach that stage, it is most likely that it will be, a friendly (hopefully) Superintelligence, which will control humans in a kind of parent-child relationship
- Comprehensiveness. Such control should apply to all AI products (robots, AI-chips, AI-controlled infrastructure, weapons and military equipment, satellites and rockets, economy, financial sector, etc.), as well as services, (algorithms, AI software, brain controlling networks, brain and body implants, neuronal nets, visual and audio equipment, etc.). It should also extend beyond our planet, since very soon we shall have the most advanced AI, the first people on the Moon, and in the next decade, on Mars.
- Ultimate control. No country should be exempt from following the rules of such a global AI-Controlling Agency. This is almost certainly not going to happen, e.g. not in China, Russia, N. Korea, Iran etc. but nevertheless such laws should be in place, like the UN Charter of Human Rights, which only some countries observe.
- Partnership with a mature AI. Such legislation should envisage that a maturing AI will be taught the Universal Values of Humanity and be driven not by goals (apart for the lowest level robots) but by human preferences, keeping the AI agent always slightly uncertain about the ultimate goal of a human. Of course, these values will only be Universal, if Humanity act as oneness, as a single federation. Only then, sometime in the future, humans, although being far less intelligent than Superintelligence, will hopefully be not outsmarted, because that would not be the preference of the Superintelligence. (This is almost an idealistic vision, but unless it materializes, human species may lose the ultimate control over its destiny).
- Reform democracy. The prime goal here should be to revise the existing Charters of Human Rights, such as the UN Declaration of Human Rights and the European Convention on Human Rights, signed up by 47 European Countries, and use it as a basis for the Universal Values of Humanity, on which a new Charter of Human Rights and Responsibilities could be built. The prime use for these new values, agreed by the majority of the states (but not all), should be in the development and ‘education’ of the maturing AI. Equally important will be the use of the new values as the foundation for the reform of democracy, resulting in the revision of the states’ constitutions, including the new EU Constitution. Such constitutions should cover new forms of government, consensus politics and the creation of Citizen’s Assemblies, as one of the key differentiators in taking main political decisions in the future
- Create a European Federation (EF). This should be a very shallow (Minimal state, or miniFed type) federation giving current members even more decision making then they have right now, apart from defence, foreign affairs and the common currency areas. This may come earlier than a deep reform of democracy, if some nearly catastrophic events occur. Creation of the EF is necessary to evolve it progressively into the future Human Federation, which would include all nations, and further re-enforce the control of AI development as it embraces all areas of human activity.
- Global Wealth Transfer. Create Global Wealth Redistribution Fund (GWRF) – a comprehensive, 20-year programme of substantial economic help for the new members of the European Federation – the full proposal in my previous book Who could save Humanity from Superintelligence? Apart from purely ethical reasons, this will also become a shield against other catastrophic risks. These include severe draught, mass migration and political disorder, which when combined with other catastrophic events, such as a more aggressive pandemic than the Covid-19, it may become an existential risk, leading to Humanity’s extinction. Therefore, the creation of GWRF should really happen even before the European Federation has been set up.
- Create a Human Federation:
- Expand the European Federation by inviting into the EF Association Zone, countries, which meet only three conditions: agree a complete demilitarization, have internationally recognized borders and accept the European Federation’s Universal Values of Humanity. Such a rapid expansion would then create a domino effect until the time when all nations would join. At the same time, the AI legislative framework would naturally cover more and more countries.
- Demilitarize the world following the Superpowers’ acceptance that there can be no winner in any global war, even if assisted by presumably the most advanced weaponized AI system, because of the consequences of the AI Supremacists’ Dilemma. This would be the moment when all Superpowers’ own weaponized AI, together with other conventional and nuclear weapons, would be disabled. Once all Superpowers have joined the European Federation, all scientific discoveries and progress in the AI development would be fused to assist in creating a mature, friendly, singular Superintelligence. The age of human co-existence with a friendly Superintelligence will then begin.
It is quite likely that unpredicted events, such as the Covid-19 pandemic in the early 2020, may change that order from ‘what should be carried out’ to ‘what’s possible’ or ‘what must be dome immediately’ at the time. The final question is how you and I could contribute to the most desired outcome. You will find my view on that here.
Tony Czarnecki, Sustensis