Email Subscriptions

Artificial Intelligence

Will Artificial Intelligence Surpass Human Intelligence? - A Viewpoint

Applications of AI are ever-increasing, making one wonder if it is going to erase the importance of human skills and experiences. Artificial Intelligence (AI) is one of the most disruptive technologies of recent times.

Insights

AI has proved to give better outputs than humans in some areas.
Artificial Intelligence is a machine's ability to perform cognitive functions like perceiving, learning, reasoning, etc., usually associated with human minds. Alexa, Siri, Chat GPT, Sophie - the humanoid robot, recommendation algorithms in Amazon or Netflix, Facebook (friends’ recommendations), and Tesla self-driven cars are all examples of AI.
Will Artificial Intelligence surpass Human Intelligence? Let us find out.

Introduction

Artificial Intelligence is very much a part of our day-to-day lives. We ask Siri or Alexa to play our favorite music. We tag the faces of our friends in our social media photos. We see personal recommendations during our online shopping experiences and come to know of products that suit our needs. We use autocorrect to save time when typing and prevent spelling mistakes. We use chatbots during our interaction with customer services. ChatGPT has been creating waves since its release. We use autonomous vacuum cleaners, climate control ACs, and so on. The reach of AI is very wide.

Human beings are born with natural intelligence that starts learning and thinking very early. We analyze problems and then solve them. We have emotions and can empathize and express our feelings. We reason, we question, we are affected by the surrounding situations, and we understand the circumstances, and we can change our decisions.

A robot cannot take care of a child with the same affection as the mother. When we accidentally pick up a hot object, we immediately throw it away, whereas AI is not capable of such a reaction. When a dog is hit by a car, we sympathize and try to help, whereas a robot cannot do that. When an orangutan's baby is separated from the mother because of poaching, we feel sad and decide to support anti-poaching campaigns, whereas AI cannot decide anything for which it is not trained or programmed.

Will AI be able to match humans in everything they do?

Key Drivers of the Artificial Intelligence Surge

AI progress is possible due to several factors, which includes:

The emergence of big data
Surge in data availability
Increase in AI investments
Increase in the computational power of machines leading to deep learning with networks of neurons composed of layers of processing units

Some Well-known Names in the AI Market

Some Well-known Names in the AI Market

The Current State of Artificial Intelligence

Artificial intelligence is categorized into 3 types:

Narrow Artificial Intelligence (ANI): Also referred to as Weak AI, it is specialized in a particular area and can outperform humans in a specific task. All examples shared earlier, like Tesla, Siri, and Alexa, are part of Narrow AI. It is dominated by big data, deep learning, and statistical approaches.

General Artificial Intelligence (AGI) or Strong AI is when it can perform all intellectual tasks at the same accuracy level as a human.

Artificial Super Intelligence (ASI) is when AI will be superior to human cognition in all aspects. AI will surpass human intelligence at this stage.

The Three Types of Artificial Intelligence

Ongoing research investigating brain emulation involves:

Cognitive neuroscience or an attempt to understand the workings of the human brain
Thought experiment demonstration in theory, the possibility of creating machines with human capabilities
Create machines exhibiting human-like intelligence called Strong AI or AGI

Artificial Narrow Intelligence is the AI currently in use and is rated as weak. General Intelligence and Superintelligence depend on the pace of AI research, breakthroughs in human cognition understanding, and advancements in hardware and software systems. General Intelligence is a theoretical concept at present. Technological singularity is a hypothetical future point in time where Superintelligence systems recursively self-improve and design themselves, resulting in unforeseeable changes to human civilization.

The Waves of Artificial Intelligence

According to DARPA (US’ Defense Advanced Research Projects Agency), AI development comes in 3 waves:

The Waves of Artificial Intelligence

First and Second-wave AI are Narrow Intelligence systems. Most of the current AI systems belong to this wave. With recent breakthroughs in Deep Learning and Generative AI, the third wave is just around the corner, but even the best current AI system is not considered General Intelligence. Even ChatGPT is not considered a third-wave AI.

Third Wave AI will overcome the black box nature of today's AI and will be able to recognize the picture of an apple and explain why it’s an apple and how it arrived at that conclusion. It will overcome the limitations of today’s machine learning systems, which function well in the majority of cases but can fail when presented with a case that doesn’t fit its training model.

The Artificial Intelligence vs Human Intelligence Debate

AI is better than humans in performing specific pre-defined tasks on which it is trained. AI is unmatched by humans when it comes to Human Behavior - creativity, general wisdom, and problem-solving. Human consciousness is far superior to AI. We are still a very long way from affordable forms of AI, which are as

Recent Breakthroughs in Artificial Intelligence

AI history was dominated by the study of neural networks in the initial decades; machine learning applications began to emerge next, and then came Deep Learning. Recent breakthroughs in Generative AI have disrupted the approach to content creation. Generative AI systems are a part of Artificial Intelligence and a sub-field of Machine Learning and Deep Learning.

The breakthrough transformer model architecture is the building block behind the Large Language Models, which use huge internet datasets to create textual content. Transformer-based models such as Generative Pre-Trained (GPT) language models are a type of Generative AI model.

Before the invention of transformers, deep learning models like Convolutional Neural Networks (CNN) and Recurrent Neural Networks (RNN) were popular. CNN was successful for image recognition and RNN for Natural Language Processing to some extent. AI systems fall under different branches, some of which are given below:

Branches of AI

Automation Potential

Automation Potential is determined by Technical Feasibility, Cost to Automate, Relative Scarcity (if labor is more in supply and cheaper), Benefits offered beyond labor costs - skills (better quality work, fewer errors), Regulatory and social-acceptance issues, acceptance of machines.

Areas Where AI is Likely to Replace Humans

Activities where the technical potential for automation is high: Predictable physical work/activities in sectors like manufacturing, food service and accommodations, and retail. These are mostly low-skilled jobs. Activities in middle-skilled jobs in data collection and data processing also have high rates of automation.

Activities where technical potential for automation is medium: Activities in sectors like Financial Services, Insurance providers that involve large data collection and processing. Operating machines in unpredictable environments like farming and construction have medium potential for automation.

Some Examples of AI Applications are Given Below:

Areas Where AI is Likely to Replace Humans

Areas Where Artificial Intelligence is Least Likely to Replace Humans

Managing and mentoring people, decision-making expertise, planning, or creative work are hardest to replace with AI. Characterized as knowledge work, these highly skilled/creative activities can include coding software, creating recipes or menus, writing promotional materials, etc.

AI excels at well-defined activities, such as optimizing travel routes, but humans still need to define the goals and interpret results. The solution needs commonsense checks.

Activities in sectors like healthcare and teaching have a low impact on AI.

Some Examples Where AI is Least Likely to Replace Humans are Given Below:

Areas Where Artificial Intelligence is Least Likely to Replace Humans

What Makes Artificial Intelligence Technologies Useful?

Critical benefits offered by AI include 1) automation of tedious processes without fatigue, 24x7 availability and reducing human errors, 2) enhance products or services by improving efficiency, reducing cost, and better performance, 3) efficient analysis and accuracy – faster and accurate analysis, predictions, recommendations, data interpretation and data-driven decisions.

Where do Current Artificial Intelligence Systems Fall Short?

AI can perform specific tasks with precision and better than humans. But still, there are a lot of areas in AI that need improvement.

Where do Current Artificial Intelligence Systems Fall Short?

Is Artificial Intelligence a Risk for Humans?

Unemployment concerns: Repetitive or data-driven jobs, manual jobs that can be automated, will be replaced by AI. This has led to widespread concerns about AI replacing human jobs. Low-skilled jobs are more at risk of being replaced by AI, and a switch in occupations may be needed.
Workplace changes: An increase in collaboration with machines will change the nature of jobs. Eg: A data analyst will have to learn to use an AI system. AI adoption in the workplace will lead to a re-evaluation of skills and competencies and an increase in digital education. Average wages might be impacted.
Shift in skills: AI will result in a shift in skills in the workplace. Education requirements will also change accordingly. Physical and manual skills and basic cognitive skills will see a decline. Eg: data collection, data processing, and predictable manual work. Higher cognitive skills, social and emotional skills, and technological skills will see a rise. Eg: creativity, critical thinking, and complex information processing.
Social impact: Economic disparity may rise as one section of society will benefit from AI, and another may face job security issues and job replacements. Dependence on machines may lead to a lack of creativity, loss of human touch, and reduced social skills. The education system should evolve to provide digital education.
Misuse & misinformation: Chances of misusing AI are there where the data can be manipulated, for example, social media manipulation or creating deepfakes.
Technological singularity: AI experts and scientists warn of dangers of singularity and believe this will result in irreversible changes to human civilization.

According to a report by Goldman Sachs, AI could replace the equivalent of 300 million full-time jobs. According to the McKinsey report, 400-800 million individuals globally could be displaced in the next 15 years due to automation. Out of the 60 percent of occupations, nearly 30 percent of activities can be automated. Out of the displaced, 75-375 million individuals may need to switch occupations. However, AI will fuel economic growth and productivity boost and also create new jobs, offsetting the lost jobs.

How Does Automation Impact Statistics Look Like?

Automations Replacing Jobs Statistics Expected by 2030

The global workplace would be impacted by multiple factors, like the cost of AI deployment, labor feasibility, and the pace at which AI adoption is done. The AI adoption would defer based on the countries and market dynamics, labor supply quality-quantity, wages, etc. Though there are so many dependencies, the major ones are technical, and their importance is given to AI automation.

Automations Impacting Jobs Statistics

Automations Impacting Jobs Statistics

Technical automation potential - ~50% current work activities that can be automated. As per PWC, jobs will be created in some sectors, and there will be job loss in other sectors by 2037, as shown in the below graph.

Change in Job Market Because of AI by 2037

How Can we Prepare for the AI Revolution?

Organizations should consider the impact of AI on their strategy and investigate how it can be applied to business problems or to gain a competitive advantage and prepare for the AI revolution. The McKinsey report highlights the need for urgent action from business leaders and policymakers to prepare for the AI revolution.

Changes to the education systems for digital education and re-skilling and training for workplace changes.
Companies need to become more agile and nonhierarchical to redesign work and the workplace.
Productivity growth and job creation can be driven by entrepreneurship and new business formation.
Unlocking investment and demand and embracing AI for strong economic and productive growth. Job growth will be fueled by strong economic growth.
Strong economic growth can help support incomes – ideas like social safety nets, universal basic income, minimum-wage policies, etc.
Through incentives, companies can be encouraged to invest in human capital.
Plan transition support and safety nets for affected jobs, mobility support, and temporary transfers. Digital talent platforms can help with market dynamism.
Embracing AI and automation, safely factoring in data security, privacy, judicious use, and overcoming bias.

Conclusion: Can AI Do Without the “Human” Factor?

AI can perform repetitive tasks with great precision and make data-driven decisions, and hence, it is likely going to replace humans in low-skill areas or predictable physical work, and it will augment human capabilities in other areas requiring medium-skills. Knowledge-driven, highly skilled areas are least likely to be replaced. The potential for automation varies across activities and sectors. AI can result in productivity boost and economic growth. New jobs will also be created. AI has entered a phase where it is generative and creative. AI can make accurate decisions using predictions, classifications, and clustering. With AI, tasks or activities previously deemed impossible have been possible. However, this is not the first step of AI domination. Similar apprehensions were expressed when computers became mainstream, but we are not yet their slaves (Despite our obsession with computers and smartphones).

Even the best AI system has not reached General Intelligence and has areas for improvement. Currently, AI systems are considered Narrow Intelligence. If AI must match human intelligence, the systems must advance to the General Intelligence stage. Further, for AI to exceed human intellect, it must reach Super Intelligence. This depends on the pace of AI research, breakthroughs in human cognition understanding, and advancement in hardware and software. There are arguments and counterarguments from the scientific community that AI may never model or replicate the human brain. General Intelligence is a theoretical concept at present and may be a reality someday, but it will be known only in the future. The states where Artificial Intelligence will surpass Human Intelligence, i.e., Artificial Superintelligence and Technological Singularity, are hypothetical states of AI.

AI is making progress only because of human contributions. Humans have an irreplaceable role in the advancement of AI, across all stages from research and development to supervision of its operation and usage, to taking responsibility to avoid misuse. In other words, there is no Artificial Intelligence without Human Intelligence. AI should be seen as a productivity improvement tool which is used to augment human capabilities rather than as a complete replacement for humans.