How do we define AI
Artificial intelligence is intelligence exhibited by machines. The majority of the AI research is focused in the field of computer science, where AI is perceived as “intelligent agents” where the goal is to “describe and build agents that receive percepts from the environment and perform actions” (Russell, 2015). The intelligent agent described here is a rational agent which “for each possible percept sequence, a rational agent should select an action that is expected to maximize its performance measure, given the evidence provided by the percept sequence and whatever built-in knowledge the agent has”. (Russell, 2015)
Already in the Dartmouth tradition, the learning capabilities of intelligent agents were highlighted. We believe that learning is an essential characteristic of AI. It is the key feature separating AI agents from traditional software. Although AI remain as a varied collection of technologies, the learning capability is the feature, which gives AI agents the revolutionary power.
In the development of intelligence above human-level, we can separate five different approaches: artificial intelligence, whole brain emulation, biological cognition, brain-computer interfaces and networks & organizations (Bostrom, 2014). In our discussions at AI Society, we will limit our discussions to artificial intelligence and networks & organizations -approaches. Former refers to programming-based approach and the latter focuses on developing collective intelligence and connecting that through defined networks.
Not replicating human intelligence - necessarily
The relation to human intelligence is not relevant in our perspective. Discussions on logic, processes and intelligent behavior of AI are too centered on humans. Intelligence is not necessarily limited to the intelligence expressed by humans and being intelligent does not mean behaving like a human.
We believe in the goal-oriented approach: if artificial intelligence is able to produce similar or better results than human intelligence, that is the most important reasoning and premises for the discussions of AI impacts.
We are aware that this limits the question of consciousness out of our scope. Artificial consciousness and its possibility will be a major topic in many fields, but we need more scientific progress before the discussion will turn meaningful. While we enjoy discussing utopian topics, we prefer to studying more concrete topics that can help us understand AI in a shorter term.
Not only hype
Recent developments in the areas of game-playing AIs justifies that this time the fuzz around AI isn't unfounded. We don't know if the AI development will reach the human-level, but obtained capabilities in areas of learning, strategic thinking, and even creativity prove important breakthroughs. It seems that current technology level is enough in order to have a radical impact on our society.
We believe that AI effect doesn't apply anymore as the new technologies are capable of learning and improving (as our definition describes). ”AI effect” means a situation where something is no longer perceived as artificial intelligence because we have managed to create it. AI effect isn't the most exact concept, but if its disability will be verified, it will have a great influence on how we see the development of AI.
Focus phenomena existing now or in the near future
The scope of our group is to study and discuss the impacts artificial intelligence has on society. We will be studying observable phenomena and focusing on next few decades.
We are purely a multidisciplinary community. We connect researchers and professionals from the fields of social science, medicine, philosophy, economics and business, technology, law, education, and linguistics. We will participate in the discussions in this fields, but also combine the knowledge and views of our professionals for the broader questions of AI impacts field.