The tech sphere is an all-encompassing umbrella term that includes a wide, diverse and seemingly constantly expanding number of fields, each of them promising to become the next big trend that will take the market by storm. But as has often been the case in technology, today’s hot topic can quickly become relegated to the sidelines and replaced by a new buzzword that entrepreneurs are quick to introduce into their discourse.
After introducing the revolutionary AI staking program through which we offer every human being the opportunity to govern AI applications from the Humans.ai ecosystem, we propose a brief introduction to the technology that is considered one of mankind’s greatest inventions: Artificial Intelligence.
Although this is generally the case for most technologies that manage to get into the spotlight, there are exceptions to this rule, and one notable example is artificial intelligence which managed to remain a constant presence in our collective consciousness.
Even if our fascination with AI has remained constant over the decades, the concept itself still seems alien to some and shrouded in mystery for others. Humans.ai aims to take our fascination with AI and turn it into reality, into something that everyone can interact with, and more importantly, trust. To achieve its goal, Humans.ai will leverage blockchain technology to put a human behind every AI decision to enable enhanced traceability, accountability, and ownership.
What exactly is AI?
Although the question is straightforward, it’s not as easy as it seems. The concept of AI can conjure up different things depending on the person asked. For some, AI invokes the dry and bone-chilling voice of HAL 9000. Others may say that it has something to do with Asimov’s Laws of Robotics, or if you ask a computer science student, he may say that it’s an algorithm that can produce results without requiring instructions. What makes this question tricky to answer is the fact that all of the above answers are to some extent true. AI can mean many things, depending on what the end goal is. Also, the absence of an officially recognized definition doesn’t particularly help our plight. As the field is constantly evolving, new concepts are introduced while others are discarded as no longer being considered to be within the boundaries of AI.
By analyzing the word itself we can better understand what the concept implies. Artificial Intelligence is an open compound word formed of the word artificial which defines something that is man-made and intelligence which refers to thinking power. Put together, artificial intelligence denotes a man-made thinking power, an intelligent entity created by humans, capable of performing certain tasks with or without instructions.
Regardless of how we want to define it, AI has become one of the most complex tools employed by companies and people on a daily basis. The average person uses AI when he asks Siri or Alexa to cancel an appointment or when he is using apps such as Google Maps or Waze to find the quickest route in traffic. On the other hand, companies are looking into AI to make better market predictions, automate repetitive tasks, make advertising more intelligent, detect patterns in data sets as well as bridge the gap between science fiction and reality with new technologies like self-driving cars and synthetic media.
A brief history of AI
Unsurprisingly, our fascination with intelligent inanimate objects predates the concept of artificial intelligence by a couple of thousands of years. Multiple ancient cultures around the world have various depictions of inanimate objects endowed with intelligence. The Greek god Hephaistos, the god of blacksmiths, metallurgy and craftsmen, is depicted in various myths as the creator of robot-like servants capable of acting of their own accord. The most famous creation of the Greek god is Talos, a giant bronze automaton tasked by the gods to guard the island of Crete.
It wasn’t until the advent of modern computing in the 20th century that a generation of scientists, mathematicians, and philosophers spurred by legends and science fiction literature started to lay the groundwork for this new field of study. One of the first people to explore the mathematical possibility of artificial intelligence was the British mathematician and World War II code-breaker Alan Turing, who argued that humans rely on available information and intrinsic reasoning to make decisions and solve problems, a pattern that he believed could also be applied to machines. In his seminal paper from 1950 “Computing Machinery and Intelligence”, Turing outlined how to build intelligent machines and how to test their intelligence through the imitation game, which later came to be known as the Turing test.
Another pivotal moment for the field of AI is the year 1956, which is often cited as the birth year of the modern field of AI. During a summer conference at Dartmouth College sponsored by the Defense Advanced Research Projects Agency (DARPA) the term artificial intelligence was coined by John McCarthy, who is considered the father of AI. McCarthy defined artificial intelligence as “the science and engineering of making intelligent machines, especially intelligent computer programs”. During the conference, Allen Newell, a computer scientist, and Herbert A. Simon, an economist, political scientist and cognitive psychologist, unveiled their Logic Theorist, a computer program that was capable of proving mathematical theorems, which is widely regarded as the first AI program.
Fast forward to the present day, and we can observe that the tremendous leap in computational power and the massive amount of data that is collected on a daily basis has spurred a period of renaissance for AI, which is no longer relegated to the pages of science fiction novels, becoming a constant presence in our lives.
Types of AI
Artificial intelligence is a multidisciplinary field with a seemingly inexhaustible pool of applications that can be categorized based on their capabilities and on their functionalities.
Based on capability, artificial intelligence falls into three broad categories:
Artificial Narrow Intelligence (ANI)
Artificial General Intelligence (AGI)
Artificial Super Intelligence (ASI)
Artificial Narrow Intelligence
If we were to draw a roadmap of the evolution of the field of artificial intelligence, we would be positioned in the Artificial Narrow Intelligence zone. Although this may come as a surprise, everything we have managed to build so far in terms of artificial intelligence falls within the artificial narrow intelligence category which means that we have just started to scratch the surface of this field. Artificial Narrow Intelligence systems are designed to perform a specific task autonomously using human-like capabilities. By definition, ANI systems have narrow capabilities like finding the quickest route in traffic, recommending products based on a user’s search history or predicting a weather forecast. As such, ANI systems work well in controlled environments at tasks they have been programmed to do but are unable to do anything that exceeds their capabilities. The most widely known examples of ANI systems are Siri, Alexa, and Cortana.
Artificial General Intelligence
The next step in the evolution of AI systems is the pursuit and creation of systems that are on par with humans in terms of their ability to learn, perceive, understand and reason. Such systems fall into the Artificial General Intelligence category which so far has eluded researchers all over the world. Some researchers speculate that an AGI system would need to be composed of thousands of ANI systems working in tandem in order to mimic human-level reasoning. At the moment of writing, the biggest obstacle in the way of developing true AGI systems stems from the tremendous computing power necessary to simulate the complexity and interconnectedness of the neural activity that takes place in the human brain.
Artificial Super Intelligence
While Artificial General Intelligence seems like a distant but achievable proposition, Artificial Super Intelligence is at the moment well-rooted in the realm of science fiction. As the name implies, ASI systems would be able to far surpass humans in terms of cognitive functions, being able not only to mimic the multi-faceted intelligence of human beings but overwhelmingly surpassing humans with its enhanced memory, fast data processing and analysis and nanosecond decision-making capabilities. Often referred to as the last invention of mankind, it is speculated that ASI systems would trigger an uncontrollable and irreversible period of technological growth dubbed the singularity that could eventually lead to unforeseeable changes to human civilization.
From a utility perspective, artificial intelligence systems can be split into four distinct categories:
Reactive
Limited Memory
Theory of Mind
Self Aware
Reactive machines
The most basic type of artificial intelligence system is represented by purely reactive machines that are unable to store memories and learn from past experiences to improve decision making. This type of system is highly specialized to solve a particular task by perceiving the world directly and calculating the best possible solution to the situation it is presented with. Probably the most famous example of a reactive machine is Deep Blue, IBM’s chess-playing supercomputer which managed to beat international chess grandmaster Garry Kasparov in 1997. Another example is Google’s AlphaGo which managed to beat a number of top human Go experts.
Limited Memory
Limited memory machines possess the capabilities of reactive machines, with the added benefit of storing and processing information in order to enhance their decision-making capabilities. A large segment of AI implementations fall within this category. Deep learning and machine learning systems are trained following this paradigm. They work by ingesting large quantities of data to detect patterns. We, humans, excel at recognizing patterns, but what seems easy for us is actually very hard for computer systems. For example, image recognition AI is trained using thousands of pictures to learn how to label the information it receives, a task that can be easily done by a child with a sufficiently long attention span.
A good example of applied limited memory AI can be found in self-driving vehicles that are preprogrammed with a representation of the world composed of streets, traffic signs and so on. Besides the information stored within their memory, they also observe the world around them and store relevant information such as traffic speed and weather conditions to adapt to current circumstances and navigate safely.
Theory of Mind
The last two types of AI systems fall under the work in progress category, existing mostly in the form of research papers and theoretical approaches. An entity is considered to have a theory of mind if it is able to understand, read and observe emotions that are not directly observable, and contextualize and adapt its actions based on its observations. A theory of mind level of AI is a system capable of understanding humans by discerning emotions, beliefs and thought processes and changing its behaviour accordingly.
Self Aware
Self-aware AI is the next step that follows the theory of mind concept, delving deep into philosophical concepts like consciousness and self-awareness. Often seen as the holy grail of AI researchers, a self-aware AI will not only be able to understand humans on a much deeper level, but it will become an entirely new entity, with its own convictions, beliefs and emotions that will blur the lines between machine and human.
AI in practice
Over the past decade, AI has slowly crept into every corner of the business world, slowly becoming a constant presence in our lives regardless of whether we are aware of it or not.
Web searches: AI systems take the vast input of data generated by web searches and analyze it in order to provide the most relevant search results or suggestions.
Online shopping: artificial intelligence enabled online retailers to create personalized product recommendations for users based on their previous searches and purchases.
Digital personal assistants: smart devices are now integrated with AI-based personal assistants that help users go about their daily routines.
Machine translations: language translation software like Google Translate have come a long way over the years. Relying on an artificial neural network, Google’s software uses the millions of examples that are typed daily to improve the quality of translations.
Smart homes and ambient experience systems: our homes are now able to analyze and learn from our behaviour. This way sensors integrated into thermostats can help reduce energy wastage by powering down when they aren’t home.
Cybersecurity: intelligent cybersecurity systems rely on AI to analyze cybersecurity incidents, training them to recognize patterns and backtrack attacks.
Healthcare: researchers are studying how to leverage AI systems to analyze large quantities of health data in an attempt to discover patterns that could lead to the discovery of more efficient treatments. In the long run, AI systems in healthcare may help improve the accuracy of individual patient diagnostics.
AI by Humans.ai
Humans.ai is blending blockchain technology with artificial intelligence to create AI with a heart. No, we’re not creating the tinman from the wizard of OZ, at least not yet. AI with a heart is our take on artificial intelligence through which we want to put a human being behind every AI created in our ecosystem.
To achieve our goal, we patented our unique Proof of Human, a new blockchain consensus that relies on humans to leverage their biometric data to prove that a specific AI is still under close biological supervision. Our approach to the new AI revolution is centred around humans, giving them full control over the governance and management of any AI, essentially making sure that the objective of the AI is aligned with the human’s objective.
