Stages of AI Development
As previously mentioned, there are two different types of AI that can be categorised. However, there is a 7-step staged process that can help determine where that AI is, in its journey.
STAGE 1 – RULES SYSTEM – A rules-based AI is what we’ve been talking about when it comes to virtual assistants and face scanners. These AIs can only operate within their set parameters.
STAGE 2 – CONTEXT AWARENESS – When an AI is capable of putting something into context, it moves up a stage. These AIs can be trained to respond to queries based on human knowledge and experience. They can also be updated and modernised when new queries are made, that were previously unthought of. Common contextual awareness AIs are things like chatbots and purchase recommenders.
STAGE 3 – DOMAINS – A rules-based AI is somewhat similar to that of a domain master, in that it can do a wide range of things that it has been programmed to do. The difference is the level of intelligence, ability, and that these two AIs have. An AI that masters a specific domain, will be given as much knowledge as possible in that specific area. Whilst they might be able to carry out fewer tasks, the domain masters will be able to carry out one or a few tasks extremely well, almost, if not exceeding that of human ability. An example of a domain-mastered AI is cancer diagnosis machines that are able to scan the body for cancerous cells.
STAGE 4 – REASONING – This is where our knowledge limits us to. One of the main drawbacks of AIs is that they lack the ability to understand reason, circumstance, intentions and logic. However, research is being done on an algorithm that could potentially allow AI to gain these traits. Although there have been breakthroughs in recent years, this type of AI will not be put into use until there is much more knowledge surrounding the area.
STAGE 5 – CONSCIOUSNESS – As we’ve discussed, the cognitive ability of a human brain is almost impossible to reproduce. The cognitive ability in question involves thinking abstractly, solving complex problems and learning quickly, and from experience. It will be a long time before this technology is introduced, and that is of course if the technology itself is actually possible to produce. At this level, AIs are known as General AIs.
STAGE 6 – SUPER INTELLIGENCE – Artificial Super Intelligence, also known as ASI, is where the threshold is, wherein AIs will outsmart the smartest human domains. The biggest problems to mankind could potentially be solved by ASIs, however, this would also result in AIs outsmarting the humans that created them. This is the point where some people place their fears and doubts in AI, notably the people who claim to the science-fiction outcome that humanity be enslaved by ASIs in the future.
STAGE 7 – SINGULARITY – Research has shown that AI singularity could be met by 2045, provided all other stages beforehand are met with relative ease. AI singularity could result in humans and AI combining together to form a connection, almost like a living internet. Ideas and knowledge could be shared with just a thought from the comfort of your own bedroom, whilst we could potentially connect with other lifeforms to like animals and plants. In order to achieve this though, a lot of groundwork will have to be laid beforehand, which as Fujitsu have found out, will take a long time.
There’s no denying that AI will become engrained within our lifestyle, even more so than it already is. Tech companies around the world are implementing more and more AI as we speak, but that is not a reason to be relaxed about it. AI could be potentially dangerous, to find out how, “The Future of Work” podcast can explain.