May 2020
Does the development of artificial general intelligence leave most of us at the station?
Artificial General Intelligence are systems, like humans, that has the ability to reason across many domains and that can use her knowledge into unknown areas. From a possibility point of view AGI can revolutionise research of medicine, production of medicine, production of food and green technologies. If we are successful, we might be dealing with climate threat, extreme poverty and pandemic.
Max Tegmark , co -founder of Future of Life Institute, said, “I’m optimistic that we can create an inspiring high-tech future with AI, as long as we win the race between the growing power of AI and the wisdom with which the manage it.
Sometimes doomsday prophecies are written about AI and that the consequences will lead to the downfall of Earth. In fact, is that nobody knows, many have opinions, in itself a good sign.
Today's AI, compared to AGI, is still fairly simple but with great capabilities in limited areas. No one knows how an AI that is as smart or smarter than humans will affect our lives. Will it be beneficial or harmful? How should we build it so that it is safe? Can we create security mechanisms that we incorporate during the ongoing development?
However, there are some risks where one is to become number one in the race be the first to have developed an AGI. This can have huge negative consequences if ethical or other views and if the need for safety mechanisms is underestimated, or even ignored. Another risk is that in our excitement to automate every work that is monotone or boring we can forget to take long-term consequences in consideration. A third one is the centralisation of resources, both economical and competence wise, to big governmental or global actors. Actors that is vacuuming the market or the state of the smartest and most educated people. A fourth risk can be the selection of the people that will form the development teams are selected from very homogeneous groups of the population. This might lead to missing important input parameters during research, design and development of AGI. An example of that can be achieved with ordinary. “simple AI” is when AlphaGo beat the world champion in the game GO. The system does not “Think” the same way we do, it takes other kinds of risks and sees opportunity’s we would not even consider. Experts of GO found out that the system made moves that we never would have done, not even in our wildest imagination. Another example is when Google's Deepmind beat man in the game Quake III, a game with multiple participants. The system has several of instances of” Bots” that worked together. The system eliminates the “Bots” with the weakest performance after every game-. They were replaced with a clone of the best performing Bot. What happens if a system ends up in a conflict regarding its objectives and human wellbeing? How would we know that information provided by the system is what we expected, if the system makes considerations we could not even imagine? What other perspectives do we need to take into consideration to manage and control of the development of AI? How do make sure that WE decide and not the system? Several different initiatives to address the different risks highlighted is ongoing. Research in different areas have been made possible by financial contributions. Independent institutes and or organisations is engaging in the task to keep control over AI and future AGI:s. To ensure that development of AI is as safe as possible, there is a need of new institutions that can provide knowledge and help spread knowledge. Institutes that conduct research and provide advice on how to regulate the ongoing development. Provide advice to politicians, authorities and other parties involved. We need to ensure that AI is provided with mechanisms that enable them to imitate and understand human behaviour, and human interactions. Mechanisms that allow machines and people to coexist on equal terms. This will only happen with courageous leadership as well as a dedicated collaboration among necessary partakers. Will we let some big corporations, organisations or states to take the lead of the technological development of AGI? Leave them to do their race in silence? Activities that either aim to guarantee a state's security and control or create maximum profit for its shareholders. Let these participants develop systems so powerful that we can't even imagine what they could accomplish. Most of them putting their own interest first - while there is a built-in conflict between states and other contributors’ interests and individuals or groups interests.
Another interesting view is to observe the individuals that conduct the technical development. Who are they? Do they work in heterogenous or homogenous composed constellations? Which perspectives do they bring into the development? What can we small ants do then? Everyone can increase their awareness of what is happening. Make yourself aware of one of the biggest technical innovations we have seen in the history of mankind might be conducted by a few accomplices that have their own goals in focus. Use our increased awareness to engage in different fora’s. Start to lift questions towards our politicians, demand political actions and demand to be listened at. The subject is too important to be buried in an investigation that leads nowhere. The train is about to leaving the station – jump aboard and get involved. Put forward thoughts, opinions regarding every aspect what you think should be part of the game. Collaborate with others and create thoughts based on a longer perspective.
My colleague @jon has previously written that humans and AI need to coexist and that humans need to expand their consciousness to come out on top of the ever-smarter systems. Our knowledge of how things are interconnected, something that another of my colleagues @ jarle has written about, shows us that we need to consider our development from a holistic point of view. We can’t develop new technology without to take into consideration many different aspects. Learn to understand the consequences from different courses of action. Be conscious! Ulf
Коментарі