What Is General Artificial Intelligence (AI)? Definition, Challenges, and Trends

Artificial general intelligence (AGI) is defined as the intelligence of machines that allows them to comprehend, learn, and perform intellectual tasks much like humans. AGI emulates the human mind and behavior to solve any kind of complex problem. This article explains the fundamentals of AGI, the key challenges involved, and the top 10 trends in AGI advancements.

What Is General Artificial Intelligence?

Artificial general intelligence (AGI) is the intelligence of machines that allows them to comprehend, learn, and perform intellectual tasks much like humans.

With AGI, machines can emulate the human mind and behavior to solve any kind of complex problem. Being designed to have comprehensive knowledge and cognitive computing capabilities, the performance of these machines is indistinguishable from that of humans.

AGI (also referred to as strong AI or deep AI) is based on the theory of mind AI framework. Fundamentally, the theory of mind-level AI deals with training machines to learn human behavior and understand the fundamental aspects of consciousness. With such a strong AI foundation, AGI can plan, learn cognitive abilities, make judgments, handle uncertain situations, and integrate prior knowledge in decision making or improve accuracy. AGI facilitates machines to perform innovative, imaginative, and creative tasks.

Achieving strong AI has significant challenges. For example, Fujitsu has built one of the fastest supercomputers named K Computer. Although the computer broke the ten petaflops barrier, it took over 40 minutes to simulate a single second of neural activity, thereby blurring the vision for strong AI. Nevertheless, the future for artificial general intelligence looks bright as the technology can be used to mass influence society with its ability to handle complex situations, such as an economic crisis.

Multiple approaches have been tried and tested to achieve human-like intelligence. Listed below are some of the core approaches to AGI.

1. The symbolic approach

The symbolic approach refers to the use of logic networks (i.e., if-then statements) and symbols to learn and develop a comprehensive knowledge base. This knowledge base is further widened by manipulating these symbols representing the physical world’s essential aspects. The approach mimics the higher levels in the thinking of a human brain.

General AI Symbolic Approach

Theoretically, the symbolic approach can perform higher-level logic and thinking, but in reality, it lacks in learning lower-level tasks such as perception. An apt example of the symbolic approach is the CYC project started by Cycorp’s Douglas Lenat in the 1980s to advance work in AI. CYC has a vast knowledge base, a logic system, and a strong representational language.

2. The connectionist approach

The connectionist approach is a sub-symbolic approach that utilizes architectures resembling the human brain (such as neural nets) to create general intelligence. The approach expects the emergence of higher-level intelligence from lower-level sub-symbolic systems, like neural nets, which is yet to happen. Deep learning systems and convolutional neural networks such as DeepMind’s AlphaGo are good examples of the connectionist approach.

3. The hybrid approach

The hybrid approach is a blend of the connectionist and symbolic systems. The architectures leading the AGI race tend to utilize the hybrid approach, for example, the CogPrime architecture. It represents both symbolic and sub-symbolic knowledge via a single knowledge representation, which is termed as AtomSpace. The famous social humanoid robot Sophia was created by Hanson Robotics and OpenCog with the help of CogPrime, a neural architecture.

4. Whole-organism architecture

Experts believe that an actual general artificial intelligent system should possess a physical body and learn from physical interactions. While there aren’t any such systems yet, the closest one is that of Sophia — a humanoid robot that imitates human gestures and facial expressions and indulges in conversations on predefined topics. 

See More: Top 10 Python Libraries for Machine Learning

Key Challenges of Reaching the General AI Stage

Although AGI has not been realized yet, it represents a world of possibilities that can revolutionize the field of AI. Artificial general intelligence is currently marred by severe roadblocks and challenges hindering its progress.

These are the key challenges of reaching the ultimate AGI stage.

Key Challenges of Reaching the General AI Stage

1. Issues in mastering human-like capabilities

To achieve true human-level intelligence, AGI needs to master some human-like capabilities, such as:

• Sensory perception: Although deep learning systems have shown great promise in the field of computer vision, AI systems lack human-like sensory-perception capabilities. For example, trained deep learning systems still have poor color perception. This is evident in self-driving cars as they get easily fooled by small pieces of black tape or stickers on a red stop sign. A similar case is observed with sound perception. Current AI systems cannot perceive and replicate distinct human sound perception.

• Motor skills: Humans can easily retrieve any object from their pockets due to our fine motor skills. A recent development applied reinforcement learning in teaching a robotic hand to solve a Rubik’s cube. Although the demonstration is notable, it reveals the problems involved in programming robot fingers on a single hand to manipulate trivial objects such as keys.
• Natural language understanding: Humans share knowledge via books, articles, blog posts, and videos. Subsequently, when humans write, they tend to assume the reader’s general knowledge, and, as such, a lot of information is unsaid in writing.To begin with, the current AI needs to consume vast amounts of information from all knowledge sources, which is a critical task. If AI lacks the basis of common sense, it will be difficult for these systems to comprehend situations and operate in the real world.
• Problem-solving: Consider an example where a home robot has to recognize that the LED light bulb in the house is blown out and either replace it with a new one or alert someone. To carry out this task, the robot needs to have common sense as discussed above or should have the ability to simulate all permutations and combinations that determine possibilities, plausibility, and probabilities. Today’s AI lacks both common sense and simulation capability.
• Human-level creativity: AI systems can improve their intelligence on their own if they understand the vast amounts of code that humans have written, identify novel methods that can be improved, and subsequently rewrite the identified code. Although AI-based machines have been able to compose music and draw pictures, demonstrating human-level creativity for self-optimization needs further advancement of AI.
• Social & emotional connect: For AI-enabled robots to operate in the world, human interaction is inevitable. As a result, these robots will need to understand humans, facial expressions, and variations in tone to interpret real emotions. Considering the perception challenges discussed above, AI systems that are capable of empathy with an emotional connection seem a distant reality as of now.

2. Lack of working protocol

Current AI systems lack a working protocol that helps artificial intelligence or machine learning networking systems collaborate. This presents a severe technical deficiency when deploying a complete AGI system. The systems are thus forced to work as standalone models in closed, isolated environments. Such a mode of operation does not align with the complex and highly social human environment essential for AGI systems.

3. Communication gaps reduce universality

Today, AI systems face a distinct communication hurdle. Communication gaps between disparate AI systems come in the way of seamless data sharing. As a consequence, the inter-learning of machine learning models is stalled. With the impact on inter-learning, AI can fail to optimize the assigned tasks. This eventually reduces the universality of the overall AI system.

4. Lack of business alignment

For appropriate AI implementation, business executives need to take a strategic approach by setting objectives, identifying KPIs, and tracking ROI. Otherwise, it can become difficult to assess the results brought by AI and compare them to measure the success (or failure) of AI investment.

Integrating AI into existing systems is a complex process. Various parameters such as data infrastructure needs, data storage, labeling, feeding the data into the system, and others need to be considered. Currently, concerned stakeholders seem to be in the dark about all these operational parameters of AI. This hinders the overall development and achievement of business goals.

5. Lack of AGI direction

As businesses often struggle with the fundamental understanding of the AGI system, they are forced to hire a dedicated team of AI experts, which can be an expensive affair. Besides, enterprises do not have a defined AI-based plan and direction to carry out their business operations. This makes the implementation of AI platforms costly and complex. These factors contribute significantly and act as roadblocks to realizing a full-fledged AGI system.

See More: How Is AI Changing the Finance, Healthcare, HR, and Marketing Industries

10 Key Recent Trends in General AI Advancements

As AI advancements take center stage amid the COVID-19 pandemic, the development of human-like intelligence has been progressing faster than ever before. Although a complete AGI system is not a reality today, recent trends in AI may push the AGI envelope and speed up its development significantly.

Here are the top 10 AI trends that can propel advancements in AGI.

Latest Trends in General AI Advancements

1. NLP development

Natural language processing (NLP) is an AI technology that comprehends human language and significantly reduces the necessity to interact with a screen. AI-enabled devices can turn human languages into computer codes used to run applications and programs.

Recently, OpenAI released GPT-3, the most advanced NLP version to date. GPT-3 uses over 175 billion parameters to process languages. Moreover, OpenAI is also working on GPT-4, and it is expected to handle around 100 trillion parameters for comprehensive language processing. With such AI advancements, developing machines that can interact and engage with humans in a manner that is as good as real is a definite possibility.

2. Metaverse 

Metaverse has been thriving as companies and individuals explore immersive technologies to work and interact in this virtual world. According to November 2021 data from DappRadar, users spent around $106 million to buy virtual property in the metaverse, focusing on digital land, luxury yachts, and other assets.

Considering this trend, AI and ML are expected to drive metaverse forward by building a virtual world with virtual AI chatbots where users feel at home.

3. Hyper automation 

Multiple industries leverage AI and ML technologies to automate several of their processes, from robotic process automation (RPA) to intelligent business process management. Hyper automation adds an additional layer to advanced automation capabilities as it scales the automation prospects for organizations. According to an April 2021 forecast by Gartner, the hyper-automation market is expected to reach $600 billion by 2022.

4. More governance jobs

Algorithm bias can crop up from the lack of model governance. Here, AI experts are expected to pay more attention and ensure that AI/ML models do not develop biases or make bad decisions. Recently, in October 2021, Twitter admitted that its algorithm developed a bias that favored right-wing politicians and news outlets. 

In a similar incident, Amazon realized in 2015 that its algorithm for hiring employees was biased against women. This was because the algorithm scrutinized resumes from the past ten years, and since most of them were men, it was trained to be biased against women.

Examples such as these will pave the way for a rise in positions such as chief AI officer, chief AI compliance officer, and so on. With the rapid adoption of AI/ML, such instances are expected to rise in the near future.

5. Rise in low-code or no-code AI 

Today, the demand for skilled AI engineers is high. Organizations are continuously looking for engineers that can develop AI algorithms and tools to meet their business operations. Low-code and no-code AI solutions can address this issue by offering intuitive interfaces that help in creating complex systems.

Generally, low-code solutions provide drag-and-drop options, thereby easing the application building process. Moreover, NLP and language modeling technologies can also be used to provide voice-based instructions to complete complex tasks. 

6. Workforce augmentation 

The fear of AI replacing human jobs has been around for quite a while now. Factually, organizations seem to be using AI/ML models to gather and analyze data and derive insights that help in making business decisions. Businesses must have employees and AI machines working in tandem in such a scenario.

Several departments, including sales, marketing, and customer service, are already using AI/ML systems to aid their operations. However, this hasn’t reduced the potential dependence on humans. In fact, it has only increased the effectiveness of such departments. Such a trend is only expected to rise from here on.

7. Conversational AI chatbots 

Conversational chatbots refer to AI-enabled virtual assistants. They carry out natural conversations and certain rule-based operations, such as responding to queries or resetting passwords. These chatbots have replaced customer support agents, thereby considerably reducing businesses’ operational costs. With the evolving NLP landscape, conversational AI chatbots will possibly revolutionize the field of AGI in the future. 

8. More focus on AI ethics 

AI use cases have gone up significantly across industry verticals in recent times. Despite the benefits of AI technologies, the potential risks of AI cannot be ignored. As a result, the focus on AI ethics will rise over the coming years as things could turn on their head if such technologies are not used for the good.

9. AI-based hiring process

As the pandemic has already dented the hiring process, companies are now expected to use more AI/ML-based systems as the virtual world replaces the conventional physical world. Moreover, with advancing language modeling techniques and an increase in sophistication of conversational AI chatbots, employers are expected to use AI-powered tools to take care of the hiring process.

10. Quantum AI 

Although considerable progress has been made in the AI field in the last few years, quantum AI could further push AI boundaries as quantum computations could speed up ML algorithms and achieve results in a shorter time. Quantum AI could neutralize AGI obstacles as it could help create a strong knowledge base by analyzing huge volumes of data found in books, articles, blog posts, and other similar sources in minimal time.

See More: 10 Experts on the Future of Artificial Intelligence

Takeaway

The ongoing decade will be extremely crucial for the development of AGI systems. Experts believe that there is a 25% chance of developing human-level AI by 2030. Moreover, the rising inclination for robotic processes and machine algorithms, coupled with the recent data explosion and computing advancements, will offer a fertile ground for the proliferation of human-level AI platforms. It is only a matter of time before AGI systems become mainstream in this highly technological world.

Do you think AGI will be realized anytime soon? Comment below or let us know on LinkedInOpens a new window , TwitterOpens a new window , or FacebookOpens a new window . We’d love to hear from you!

MORE ON ARTIFICIAL INTELLIGENCE

• Top 10 Open Source Artificial Intelligence Software in 2021
• What Is Artificial Intelligence (AI) as a Service? Definition, Architecture, and Trends
• 10 Most Common Myths About AI
• What Is Artificial Intelligence: History, Types, Applications, Benefits, Challenges, and Future of AI
• What Are the Types of Artificial Intelligence: Narrow, General, and Super AI Explained