Scientists formulate a theoretical framework for collective intelligence, helpin

2024-05-30

Human society is able to form, maintain, and function, all with the help of intelligence. In the animal kingdom, the more clever an animal is, the more frequent its social activities become.

Some scholars have proposed the "social brain theory," which states that humans are smarter than other primates because they need to handle and maintain various social activities.

In other words, it is because of our social needs that we have become smarter and smarter.

In human society, intelligence is not isolated but is collectively owned and created by everyone. This collectively created intelligence is called "collective intelligence."

With the continuous advancement of information and communication technology (ICT), humans can more easily and at a lower cost carry out online cooperation, which also allows collective intelligence to be elevated to a higher level.The Internet, especially the World Wide Web, was designed specifically to help people work together on tasks.

The World Wide Web has enabled humanity to complete unprecedented, large-scale collaborative projects, such as Wikipedia (with its Chinese counterpart, Baidu Baike), various citizen science projects, as well as collaborative software development and tagging projects.

Advertisement

While large-scale collaboration and crowdsourcing activities are not entirely unique to the Internet, it is undeniable that the Internet has greatly facilitated the way humans collectively create knowledge, accomplish tasks, and make decisions.

Artificial Intelligence (AI), like the Internet, is rapidly changing every aspect of human life, influencing the way humans make decisions individually and collectively.

For example, the advent of intelligent devices that can learn, as well as home appliances that can converse with humans in natural language, has transformed the way humans live, work, and even socialize.Therefore, it is a matter of urgency to explore how AI, like the ICT revolution, can help human collective intelligence to reach a higher level.

At the beginning of the study of collective intelligence, people only focused on cooperation between people. However, people have started to consider AI as a member of the team, known as "hybrid intelligence."

Those who support this idea believe that if humans and AI are combined, they can make wiser decisions together and even achieve goals that cannot be achieved by a single person or a single machine.

Currently, researchers are exploring how to let humans and computers work together to make the human collective smarter than any individual or any computer.

In recent years, with the rapid development of AI technology, the role of AI in human teams will no longer be just a tool, but also a member that can participate in decision-making.AI has significant advantages over humans in many aspects, such as the ability to process vast amounts of data more quickly, identify patterns invisible to the naked eye, and provide insights based on complex algorithms.

 

Humans, on the other hand, possess abilities that AI lacks, such as intuition, complex emotions, understanding of cultural and social contexts, and ethical and moral concepts.

 

Both humans and AI have unique strengths. If these strengths can be combined, it is possible to create a new type of super intelligence.

 

This intelligence surpasses the capabilities of a single individual or a single AI, capable of solving complex problems with greater efficiency, accuracy, and creativity.

 

Therefore, researching how to integrate AI into human teams to enhance the collective intelligence of the entire group becomes extremely important.However, currently this field still lacks a comprehensive theoretical framework. Based on this, Dr. Cui Hao, who is currently a postdoctoral researcher at the University of Dublin in Ireland and his team, have written a review paper.

In the paper, the research group introduces key concepts, the latest developments, and future prospects. Cui Hao said that AI-enhanced collective intelligence can be used in the field of social media information dissemination.

Previously, during her doctoral studies, she had explored the retweet networks of Weibo's hot search topics and their evolutionary patterns over time. The authenticity of information on social media is difficult to discern, and the spread of false information may mislead public opinion and have a negative impact on society.

With the development of AI, a large number of deepfake videos have also emerged. So, how can these pieces of information be effectively identified?

It is understood that combining human judgment with AI technology can improve the accuracy and efficiency of identification.Existing research has found that this combined approach is more effective in identifying the aforementioned videos than relying solely on humans or solely on AI.

In addition, AI-enhanced collective intelligence has made preliminary progress in fields such as healthcare, education, agriculture, finance, and the environment.

So, how did Cui Hao specifically write this paper?

It is reported that when she was studying for her doctorate at the Central European University in Austria, she had already learned about the concept of collective intelligence. At that time, she thought it was an interesting but vague concept.

After coming to the University of Dublin in Ireland to start her postdoctoral research, her mentor asked her to first broadly understand the existing research in this field.However, the concept of collective intelligence is somewhat ambiguous, and over time this concept has been redefined. Moreover, different research fields even use this term to describe different content.

For this reason, Cui Hao first regards collective intelligence as a result, produced by the integration of inputs from a large number of people.

In this research direction, the concept of collective intelligence is often confused with the wisdom of the crowd. Although it sounds similar when translated into Chinese, they are still different.

The wisdom of the crowd refers to the formation of judgments by pooling personal information from a large number of people distributed in different places, usually relying on simple statistical aggregation.

In contrast, collective intelligence stems from the synergy between intelligent individuals, who work together to achieve common goals through interconnected efforts.Another research direction defines collective intelligence as a capability to solve a series of problems that change over time under complex environmental conditions.

This team-based general collective intelligence is similar to individual general intelligence, but has been extended to the team level.

Although scholars in these two research directions have different specific definitions of collective intelligence, they are essentially compatible, that is, they both believe that the collective intelligence in a team is not just the sum of the individual intelligence of the members.

In addition, network science is also one of the research directions during Cui Hao's doctoral studies. For this reason, she uses the thinking mode of complex systems and the methods of network science to explore real-world problems.

Network science can help humans better understand the complexity of social systems, such as understanding how many different factors work together to affect the behavior of the entire system.Based on this, Cui Hao began to consider whether this set of thinking could be used to explain collective intelligence. She regards the collective of humans and AI as a multi-layered network complex system, which includes three levels: cognitive, physical, and informational.

This complex system can adapt to environmental changes and continuously evolve. In complex systems, the concept of emergence can be used to describe such phenomena: new properties appear at the collective level, which do not exist at the individual level.

Under this framework, compared to individual intelligence at the micro level, collective intelligence at the macro level is a more excellent new property.

This property includes results, capabilities, and behaviors, which are compatible and encompass the existing definitions of collective intelligence.

Cui Hao stated that collective intelligence is the result of the joint action of various complex, nonlinear relationships within the system, which may include the combined influence of small-scale interactions (bottom-up) and larger-scale structures and rules (top-down).While she was writing this review paper, new review papers were also being published continuously.

 

She expressed: "This put a lot of pressure on me, to the point where I needed to find innovative points and new perspectives. When I saw a paper by a team from Carnegie Mellon University in the United States, my heart was already half cold, thinking it would overlap a lot with our paper."

 

However, after reading the aforementioned paper, she found that their focus was more on how AI affects the cognitive level to enhance collective intelligence, so there was not a large overlap in the paper.

 

When she was searching for real-world social application cases of AI-enhanced collective intelligence, she accidentally discovered the "Supermind Design" dataset.

 

The dataset contains a variety of intelligent classifications and real-world application cases in multiple fields. So, Cui Hao contacted the person in charge of "Supermind Design" and started cooperation on this basis.Ultimately, the related paper was published under the title "AI-enhanced Collective Intelligence: The State of the Art and Prospects" on arXiv[1].

Cui Hao is the first author, with Professor TAHA YASSERI from University College Dublin serving as the corresponding author.

In the future, the research group plans to delve into the specific role of AI within human teams and how it affects the overall operation of the team.

The topics they will explore include the role of AI as a team member or manager, human perceptions and attitudes towards AI members, and how the addition of AI affects team cooperation, conflict resolution, and innovation.

According to the introduction, the interaction between humans and AI contains rich value for sociological research. With the increasing popularity of AI, studying the social relationship between humans and machines will become an important field, and it is expected that this field will see significant development.Therefore, experts and scholars from various disciplines should strengthen cooperation, engage in more interdisciplinary exchanges, keep up with the latest trends, and jointly promote the development of the emerging research field of AI-enhanced collective intelligence.

Leave a comment

Submit