AI

Imagining the future and building it with AI

Future developments are now hinged upon how evolution can help develop ever-improving AI systems employing computational models.

The year 2020 will be remembered as much for the pandemic, as for being the year where technology sustained humankind and redefined how we work, live and interact. For companies, the definition of value has expanded to include how well people thrive, the impact left on the environment and adapting to rapidly changing customer needs. While we continue to get back to course post the pandemic, it is important that we adopt technologies to build for the future with a clear-eyed perspective to upend convention and replace inefficient methods and technologies with those that are increasingly advanced and interactive. There has been a clear rush to expedite digital transformation of processes towards this end and has only accelerated in the post-pandemic world.

Artificial Intelligence, or AI as its usually abbreviated, is one such technology that is powering new possibilities by enabling a digital-physical world using tools like machine learning and deep learning. To understand how AI is changing the world and how we live in it, it is important to know about its different types, current applications and how further development could radically change the future world for the better.

In terms of real-world machine applications using AI there are primarily two types, reactive or Type I machines and limited memory or Type II machines. Type I class machines can identify, calculate, predict and choose an optimal solution by running a set of pre-programmed routines. These machines perceive the world directly and lack an internal concept of the world which prevents them from learning through experiences. Examples range from industrial robots to social media monitoring tools and are ubiquitous in both the physical and digital worlds we inhabit. Type 2 or limited memory AI can look into the past for a finite period of time and use this data to arrive at a decision that is more inclusive of the external environment. Self-driving cars are examples where Type 2 AI is employed to observe traffic, adding these observations to a preprogramed representation of the car’s ‘world’ like lane markings, traffic lights and road turns and then deciding on when to change lanes so as to avoid cutting off fellow users or being hit by a nearby car.

The major limitation of Type 2 class machines is their inability to generate and store a library of experiences from which it can utilize information on a run-time basis to decide on future decisions and actions. So basically, your self-driving car can drive you to your destination but cannot understand where you may want to go next from there.

The next two categories of AI are more applicable for the machines we will build in the future: machines capable of interacting with us, understanding our expectations and adjusting their behaviour accordingly. These machines will not only form representations of the world, but also of the humans and creatures inhabiting it in order to understand that all living entities can have emotions and thoughts which in turn affect their own behaviour. This is called “theory of mind” in psychology and having this functionality is the basis of Type III AI enabled machines. Type 4 AI revolves around the concept of self-awareness and AI researchers are striving to develop machines that have full consciousness, or basically, having the ability to form representations about themselves. These machines are aware of themselves, know about their internal states, and can predict feelings of others, thus, having an extension of the “theory of mind” possessed by Type III machines.

Machines employing these advanced AI technology frameworks will be a reality sooner than we may think and be employed, for example, in the form of human-like robots performing duties in conditions not suited to human life; be that on earth or in space.

Let Everyone Talk – brands want to talk about them

Future developments are now hinged upon how evolution can help develop ever-improving AI systems employing computational models that help understand how human intelligence evolved and use that understanding to design and evolve machine applications. With AI technology now getting its due recognition amongst Industry leaders leading to an exponential adoption rate and a paradigm shift in customer preferences towards digital solutions that mirror the real world; it would be only fair to say that AI is not only rebuilding the world we live in but also reimagining our future.

This author is Neha Agrawal, Founder, MensenTock Communications and Core Mentor at NASSCOM Deep Tech Club.

 

Leave a Reply

Your email address will not be published. Required fields are marked *