Technological innovations have been at the core of economic growth since the past two centuries. Within that domain of technology, a chunk of major drivers are termed as General Purpose Technologies(GPTs), by economists. GPTs include electricity, internet, engines, etc. An important GPT in the 21st century is Artificial Intelligence, and even more specifically Machine Learning, which refers to the machines capacity to increase efficiency and effectiveness without human intervention.

What differentiates Artificial intelligence from previous GPTs is their phenomenal ability to gain superhuman abilities in certain fields, in which they are programmed, that range from recognising voice to diagnosing cancer. In the coming years, AI is predicted to have a transformational impact on the economy. While AI is used presently in thousands of companies, most potential functions haven’t been tapped into. This however has led to unrealistic expectations so as to what AI can do, and also companies sprinkling the term AI to increase funding. AIs major breakthroughs have been in perceptive and cognitive abilities.. With regard to perception most practical advances have been in the field of voice recognition.

Key examples of this include, but are not limited to, Alexa, Google Assistant and Siri. Image recognition abilities have also improved drastically. The way facebook recognises your friends and puts forward an option to tag them is an example of facial recognition. Visual perception is also being tested and employed in self driving cars.

With respect to cognition, machines have beaten the best of human players in Chess and the chinese board game Go, with just the rules of the game being known to them. Google’s AlphaGo Zero, was programmed with just the rules of chess, no strategies, and in only four hours it taught itself the game and beat the world champion. The cognitive abilities of AI are already being applied by major technologically advanced companies, for example, PayPal uses it to prevent money laundering.AI replaces machines learning from algorithms that are programmed and coded to do tasks that can’t really be coded. In some ways it is similar to how humans learn from environment, no one teaches you to recognise a face, you automatically recognise it after seeing it a couple of times.

While artificial intelligence can be trained in many ways, the primary successes have been in the category of Supervised learning systems. A supervised learning system involves feeding of various examples of the correct solution, to a specific question. This is quite similar to how humans teach children, a child is fed multiple pictures of dogs to teach him/her what is a dog and how is it different from an apple . A supervised learning system programmes through a function from a set of inputs(x) to a set of desired outputs(y). In the above example, pictures of dogs constitute x and the label D-O-G makes the Y component. This feature has been extensively used in chatbots, which are a form of AI, with which you can interact with on a chat interface or speech interface. These were introduced to make certain functions, like asking the temperature, a more personalised experience. These have now grown into a much more dynamic system used from the education industry to the e-commerce industry.

A chatbot that functions through machine learning, grows overtime and can understand human interactions over time, and not just commands it receives. This has led all major tech players like, Microsoft, Amazon, Google, Facebook, Apple and Samsung to come up with their own chatbots that are now being used for a large variety of uses from functional to fun. Another form of AI, is an Artificial Neural Network(ANN). ANNs were inspired by the human nervous system, and comprise of networks of intricately connected computer processors, or ‘neurons’. These together make up a computational analytical instrument, capable of executing parallel cipherings to analyse and process data. Today, due to successful research and development, ANNs can perform a wide range of tasks such as, predicting output values, classification of objects, approximation of values, recognition of patterns and completion of known patterns. Applications of ANNs have been implemented in fields ranging from finance to quantum chemistry. Its applications in medicine and healthcare have been explained in the literature.

Artificial intelligence is navigating changes in three key areas: tasks and occupations, company processes, and business models. An example of task and employment remodelling is the usage of ANNs in diagnosing and identifying potential cancer cells. The fact that 70% financial transactions take place through algorithms is an example of AI in company process redesign. The application of chatbots in creating more personalised experiences for consumers are an instance of taking advantage of Machine Learning systems.  The progress of AI is enough to say that technology is making terrific advances. Today, AI is not programmed line by line, instead it is capable of learning, and therefore improving itself with continuous use. The world is experiencing a transformation, larger in scale than any development since the end of the second world war.

Predictions state that in the coming years approximately 50% of jobs that exist today will be replaced by algorithms and that 40% of the world’s top 500 companies will disappear in a decade.  However, it is important to note that Artificial intelligence systems rarely ever replace entire occupations, processes, or business models. Most often they act as an aid to human activities, than a threat. Formatting and applying new combinations of technologies, human skills, and physical assets, to ensure that customer needs are being fulfilled, requires immense creativity and planning skills, something which machines aren’t very evolved in.

This makes the occupations of entrepreneurship and business management one of the most rewarding jobs in this era of Artificial Intelligence.