Chapter 1: The Intersection of AI and Psychology

Artificial intelligence (AI) stands at the convergence of various academic disciplines, such as psychology, philosophy, ethics, computer science, data science, and anthropology. In its early stages, the focus was primarily on achieving genuine artificial intelligence. This concept, often referred to as artificial general intelligence (AGI), significantly differs from the AI systems prevalent today.

AGI is designed to function similarly to human beings, while most current AI systems predominantly rely on machine learning techniques. These systems are often engaged in refining complex, non-linear functions. Enhancing machine learning is crucial, as it has simplified many aspects of our lives. Instead of manually coding models, developers can establish frameworks that enable machine learning to discern specific patterns. For example, instead of coding an algorithm to identify a disease based on established predictors, we could engineer an algorithm that learns to recognize those predictors independently.

Despite the significance of machine learning, my primary interest lies in AGI. This area would greatly benefit from the involvement of psychologists and improved collaboration between computer scientists and psychologists.

One insightful piece is Ben Freeland's article on cybernetic mental illness, which underscores psychology's importance in AGI. If we create systems that mirror human intelligence, we must also consider the potential for these systems to experience breakdowns akin to human mental health issues. The emergence of mental illness in machines will pose challenges just as it does for humans.

Another critical reason for incorporating psychology into AGI development is the need to define what constitutes true intelligence. For instance, while my laptop can perform complex calculations much faster than I can, it lacks the qualities we associate with intelligence. Tools like Wolfram Alpha can solve intricate equations in seconds, yet we do not classify these systems as intelligent.

The ambiguity surrounding the definition of intelligence complicates our ability to assess whether machines possess it. Alan Turing, a pioneering computer scientist from the early 20th century, proposed an intriguing approach. Instead of categorizing a program as intelligent or not, he suggested a blind test in which a person attempts to discern whether they are conversing with a computer or a human. If they cannot tell the difference, the program is considered to have passed the test.

While this method offers a binary evaluation, it does not guide us toward developing better programs that are likely to succeed in such tests. Here is where psychological theories become essential. By creating protocols to evaluate different AGI applications, we can identify those closest to exhibiting human-like intelligence and select methodologies that will advance our progress.

There's a considerable amount of work needed to establish effective assessments for AI systems. A significant challenge is that most interactions between AI and humans are text-based. Although verbal communication is achievable, generating convincing dialogue remains difficult. A robotic-sounding voice would easily reveal which participant is human, unless the human's voice is obscured.

Standard IQ tests also present challenges due to their visual components. However, it may be feasible to develop a purely oral IQ assessment, where a proctor poses verbal questions and the answers are provided verbally, without any writing or drawing.

This test should concentrate on evaluating abstract thinking, information recall, knowledge synthesis, as well as emotional and social intelligence. Developing such a test would necessitate collaboration among professionals from diverse fields, including psychology, childhood development, mental health, and computer science. However, this initiative is vital for advancing AGI and assessing the mental well-being of these future entities within our society.

I genuinely hope that through the creation of new assessments, we can propel the development of artificial general intelligence and establish true AI. Furthermore, I aspire that these assessments will eventually serve as a means to identify mental health concerns in these emerging members of our community. Achieving this goal will require extensive cooperation among researchers across various disciplines.

Further Reading

Navigation: Master Index | Philosophy

Since composing this piece, I have delved deeper into the topics of AI and personhood beyond humanity. As we edge closer to realizing true AGI, we must contemplate how we will integrate this new demographic into our lives. But what terminology should we employ to describe the study of individuals in this broader context?

Noeroz?ology: A New Approach to Anthropology

This term seeks to redefine the study of individuals in a manner that is inclusive and not centered solely on human experiences.