What is AI?
Artificial intelligence, or AI, is a general term (first coined in 1956 by John McCarthy) to describe a machine or computer program that can perform tasks that seem “smart” (e.g. classifying loan applicants based on their financial data and diagnosing patients based on their clinical data, etc.) A program that can perform such tasks would be said to exhibit artificial intelligence because we view such tasks as requiring “intelligence” that is normally attributed to humans.
A robot that can assemble a car in a factory or pick items from a shelf in a warehouse is also an example of an AI because it is replacing a factory or a warehouse worker. Rule-based reasoning about a domain (like medicine, finance, plant biology, etc.) can also be embedded into AI systems that were popular in ‘80s and ‘90s and known as expert systems.
In just the last year, there has been a surge of interest in generative AI, a technology that responds to user prompts in a conversational style by generating a word at a time to complete the prompt. Hence, the name generative AI. Generative AI programs like ChatGPT can keep up a normal conversation with you in plain natural language on any topic under the sun making it nearly impossible to detect that they are not human.
Over time there has been a tendency to raise the bar in terms of our expectations from AI and subject it to ever harder tests. The next holy grail is AGI or Artificial General Intelligence, an AI with an ability to do pretty much any task that humans can do, but even better. These AGIs could potentially produce successive, newer generations of even more powerful AGIs.
What does it mean for the world of business?
A business can become more efficient by exploiting AI to automate many tasks where humans were involved, thus replacing humans, and reducing labor costs enormously. Decisions made by AI are also more consistent and not subject to vagaries of human moods and whims. An AI program asked to decide on a loan application will only examine the factual data about each applicant. An automated program used to screen job applicants will again look at data about them and not be biased by extraneous factors like their gender, race, etc.
AI Chatbots are starting to replace humans in many help desk and customer support roles. Eerily humanlike AI-based assistants that are personalized to our own data and context are becoming available now and replacing human assistants. We are witnessing the dawn of a new era in which humans and machines can work together seamlessly to unleash enormous creativity and increase productivity manyfold.
How are leading organizations taking advantage of its potential to increase efficiencies, stimulate growth, and positively impact strategy?
There is a “gold rush” to exploit AI in the last year with the development of generative AI that can do pretty much everything from drafting a letter of recommendation, to ordering a pizza, or parsing your email inbox to find all messages related to a certain topic. Microsoft has recently announced a copilot software based on ChatGPT that can write very decent program code in a language like Python. The demand for software engineers is already declining. A healthcare chain is using AI to generate medical records from conversations between physicians and patients.
In just five days after its launch, ChatGPT had a user base of 1 million users. Within two months it had acquired 100 million users. Vinod Khosla, a tech industry visionary, famously remarked that in 10 years generative AI will do 80% of 80% of the jobs out there. This could produce huge increases in productivity and faster rates of economic growth. At the same time, we might see mass unemployment in a whole swath of industries and major structural changes in the economy. A new job called “prompt engineer” has emerged to describe someone who would know how best to interact with “black box” AI technology in such a way as to get it to produce the best results for questions or prompts.
There is also a lot of hype around AI that is akin to what we saw over the dot-com mania around the turn of the century. It is only when the dust settles that a realistic assessment of AI’s potential can emerge. While showcase applications may look attractive, it usually takes time to iterate through many cycles of refinement to harness any new technology so it can be put to effective use. One can expect that the new inventions in the realm of generative AI will also follow a similar pattern.
