IEEE Future Directions
Related Articles
8. The coming age of AI-Human collaboration
what iThis Megatrend is already affecting our life in many ways as well as affecting business. Think about talking to Alexa or Siri: you are actually bringing AI into your everyday life. Pick up your phone and take a picture: AI is at work when you click to determine the correct exposure given that type of image, detecting a smile on a face… Are you turning for help to Google translator to help that Japanese tourist that has asked you a question? That’s again AI at work.
In a factory plan robots are starting to interact with blue-collars and AI is supporting the interaction and collaboration. Are you in the trading floor of a stock exchange? You are most likely to use AI to support your decisions. Are you a doctor looking at a TAC and reading the report? Most likely AI had a lot to do in that report content.
What we see (and most of the time not perceive) today is just the harbinger of what is coming in the next few years. By the end of this decade (and this is what this Megatrend is about) AI will be an omnipresent companion. We have seen the starting of the market of AI on demand, AIaaS – AI as a Service – see graphic on the side, and we have seen the announcement by Google to offer access to an AI Cloud through public API – Application Programming Interface- that turns AI into – basically- an infrastructure that can be used everywhere and in any type of application.
A similar offer is available through Amazon AWS. Amazon Polly offers up to 5 million characters text to speech capabilities per month (that is over 1,500 pages of text translated into voice); Amazon SageMaker offers 250 hours per month of processing capacity to train AI models; Amazon Lex allows you to create Chatbots and having them processing 10,000 text requests or 5,000 voice request per month; Amazon Rekognition supports the analyses of 5,000 pictures and store of 1,000 face metadata per month; Amazon Comprehend lets you develop natural language understanding of up to 50,000 phrases per month and Amazon Transcribe lets you add speech to text capabilities to your application for up to 60 minutes per month: all of this for FREE! Amazing.
Notice that particularly in some areas where training is needed to create AI the process is very computationally intensive and only the big ones, with their huge computation capabilities can support this (hence the value to be able to access services like Amazon SageMaker). Having access to these resources is going to multiply the development of AI and widen its areas of application.
At the Symbiotic Autonomous Systems Initiative (now Digital Reality Initiative) we come up with a roadmap leading to a symbiotic relation among humans and machines (see first graphic) foreseeing a parallel evolution of capabilities in machines (getting more and more smarter) and in humans (getting more and more seamlessly connected with machines) that is leading, by the end of this decade, to a symbiotic life, where we, humans, will leverage seamlessly from the augmentation that machines can provide (better sensing, faster processing, specific augmentation in some areas of intelligence) and at the same time machines will be leveraging from our intelligence in a shared environment. For an in-depth discussion on these topics you can take a look at the eBook “Augmented Machines and Augmented Humans Converging on Transhumanism“.
Collaborative robots -cobots-, like Baxter (no longer in production but it signalled the starting point of a human-robot collaboration), are already a well defined area of research and more importantly are becoming industrial products.
What is interesting is that these collaborative robots were designed to be safe when operating in teams with humans. Then they evolved, through AI and machine learning, to become capable of “learning” from their interaction with humans, and, further step, to take the lead, when needed, and teach human coworkers so that the teamwork can become more and more effective.
