In these last few weeks I have been working on the preparation of an education course that will be published by IEEE on the topic of VR and AR roles in Smart Factories.
Smart Factories (has anything that can/should be labelled as “smart”) involve knowledge and being a factory an ensemble of parts, each one with a certain degree of autonomy (the smarter the factory the more autonomy is needed in its components), knowledge is distributed among them and there is a need for sharing this knowledge. This does not necessarily means to transfer knowledge from one part to another, nor to access the knowledge of a part from others. Although both of these actions (transfer/access) can occur, and might be needed, the crucial point is to make sure that the knowledge of a part “influences” the decisions/activities of the others to the point that we have an emerging knowledge of the factory that is actually greater than the knowledge of each single parts.
This is along the lines of swarm intelligence where the global intelligence demonstrated by the swarm exceeds the sum of the local intelligences (a swarm of bees is way smarter than each single bee). Swarm knowledge, likewise, is way greater than the sum of individual knowledge.
Artificial intelligence algorithms, technologies, are based on these principles (the transformers are leveraging on the knowledge of all components through their mutual influence). Unanimous AI (watch the clip) is offering this swarm knowledge/intelligence in various sectors.
Cognitive Digital Twins -CDT- is a technology, a paradigm, that makes possible to encapsulate local knowledge and operate on it in a framework of distributed knowledge. Not necessarily a CDT needs to access to another CDT knowledge nor to “acquire” it, but its behaviour can be influenced through the interaction with other CDTs, resulting in a behaviour that represents a higher level of knowledge.
CDTs can represent knowledge of a machine (they were invented by IBM to represent robots knowledge) as well as the knowledge of a human. However, as soon as we start using CDTs to represent human knowledge we open up a can of worms, raising issues such as:
- Could the CDT be used without the person’s knowledge being mirrored being aware of that?
- Would the CDT diverge, augment its knowledge through interactions to the point that the CDT is more valuable that the persons’ being mirrored?
- Who is responsible for actions resulting from the exploitation of knowledge obtained through interaction with the CDT?
- Who is tha actual owner of the CDT? In the beginning it is obvious the ownership is of the person whose knowledge is mirrored, but once the CDT expands its knowledge through interaction with a company knowledge will the company become the owner?
- How could the knowledge harvested by the CDT be transferred to the person? Could it be possible? Would it make sense (the person may just interact with the CDT to execute that knowledge without needing to transfer it…)?
This is by no means an exhaustive list (I welcome your comments and identification of other issues…), but it is sufficient to show that we are starting an uncharted path.