The Declining Critical Thinking Skills: From Artificial Intelligence to Average Intelligence

Artificial Intelligence (AI) has captivated the world with its rapid advancements, yet we might be overstating its capabilities. The term "Artificial General Intelligence" (AGI) often conjures images of machines with human-like thinking capacities. However, current AI models, including the latest language models (LLMs), might better be described as "Average Intelligence" rather than truly artificial intelligence. This shift in perception is not due to the extraordinary rise in AI's capabilities but rather a concerning decline in human critical thinking skills. As a result, AI seems more intelligent compared to the increasingly average human intellect.

What It Will Really Take to Develop AGI

Developing Artificial General Intelligence (AGI) that matches or surpasses human cognitive abilities requires several major advancements beyond the current capabilities of AI:

1. Context and Meaning: AGI needs a significant leap in understanding context and meaning rather than just processing data patterns. Current AI systems lack the ability to grasp abstract concepts and make nuanced judgments.

2. Consciousness and Self-Awareness: AGI must possess a degree of consciousness and self-awareness, enabling it to understand its existence and adapt autonomously to new, unforeseen situations. This involves breakthroughs in cognitive science and neuroscience to model human-like consciousness accurately.

3. Advanced Reasoning and Problem-Solving: Creating AGI requires developing advanced reasoning and problem-solving skills akin to human thought processes. This includes the ability to learn from minimal data, understand causality, and apply knowledge across diverse domains.

4. Emotional Intelligence: Integrating emotional intelligence is another critical aspect, as it enables AGI to interact socially and empathetically with humans, enhancing its utility and acceptance.

5. Ethical and Safety Considerations: Ethical and safety considerations must be paramount, ensuring that AGI systems are aligned with human values and can be controlled to prevent harmful outcomes. Achieving these milestones involves interdisciplinary collaboration across AI research, cognitive science, neuroscience, and ethics.

The timeline for achieving AGI is a subject of much debate among experts. According to a survey conducted by Vincent C. Muller and Nick Bostrom, 50% of participants believe AGI will occur by 2040, and 90% believe it will happen by 2075. In a more recent survey from 2022, AI experts estimated a 50% chance that high-level machine intelligence would be achieved by 2059.

Opinions vary significantly among AI researchers and entrepreneurs. For instance, in 2023, Geoffrey Hinton predicted AGI could be achieved in 5-20 years, reflecting the rapid advancements in AI technology . On the other hand, some experts believe AGI will take longer, citing the incremental nature of technological progress. Sean Holden noted that reaching AGI would require many more incremental steps, which could take several decades .

Historical over-optimism in AI predictions has led to more conservative estimates in recent years. For example, AI pioneer Herbert A. Simon predicted in 1965 that machines would be capable of doing any work a man can do within twenty years, a prediction that has not yet come to fruition . However, the rise of generative AI models has renewed optimism among some researchers and entrepreneurs. Elon Musk expects an artificial intelligence smarter than the smartest humans by 2026, while Ray Kurzweil has adjusted his prediction to 2032 .

Despite differing opinions on the timeline, there is a consensus that AGI is an inevitable outcome of continued advancements in AI. The exponential growth in processing power and memory, coupled with advancements in algorithms, suggests that machine intelligence will eventually surpass human intelligence unless there is a hard limit to AI capabilities that we have not yet encountered.

The Concept of "Average Intelligence"

Today's AI models excel at processing and generating language, performing complex computations, and even mimicking human conversation. However, they fundamentally lack true understanding, consciousness, and the ability to engage in genuine critical thinking. They function based on patterns in data rather than genuine comprehension. As Yann LeCun, Chief AI Scientist at Meta, pointed out, “a child has seen 50 times more data than the LLMs that are trained on the totality of all text that is perfectly available.” LeCun emphasizes that “4-year-olds are smarter than LLMs.” This "Average Intelligence" of AI is often mistaken for higher intelligence because it matches or exceeds the capabilities of an average person whose critical thinking skills may be on the decline. This phenomenon creates the illusion that AI is reaching or surpassing human intelligence.

Functional Illiteracy as a Metric

Functional illiteracy is a useful metric to illustrate the declining critical thinking skills in the population. It measures the percentage of people who, despite being able to read, cannot effectively understand or interpret text. This metric can be used as a proxy to assess the general decline in intellectual capabilities across different countries. According to the Programme for the International Assessment of Adult Competencies (PIAAC) by the OECD, a significant proportion of adults in developed countries exhibit low levels of literacy and numeracy skills. The PIAAC study found that many adults struggle with tasks requiring the integration and evaluation of information .

An article by APM Research Lab highlights that approximately 130 million adults in the U.S. have low literacy skills, with more than half of Americans between the ages of 16 and 74 reading below the equivalent of a sixth-grade level . Moreover, low literacy remains a serious issue across Europe, with between 20-25% of the population classified as functionally illiterate. This means that at least one in five Europeans lacks the reading and writing skills necessary for daily functioning in society, impacting areas such as education, employment, healthcare, welfare, social integration, and political participation.

In 2012, European Commissioner Androulla Vassiliou highlighted this issue, noting that over 73 million adults in the EU did not possess sufficient literacy levels to cope with daily life. Despite numerous campaigns and policies aimed at improving literacy, progress has been limited. Today, one in five 15-year-olds in the EU still has insufficient reading skills, posing significant long-term challenges for personal development and societal progress. The persistent low literacy rates risk becoming a multi-generational issue, further exacerbating social and economic inequalities across Europe.

AI Growth by Country

AI research and development are not confined to a single nation; it is a global endeavor with numerous countries making significant strides in the field. The United States leads the charge, with nearly 60% of the world's top-tier AI researchers working for American institutions and companies . The U.S. also boasts substantial private investment, with $249 billion raised to date and government spending of $3.3 billion in 2022 . Notable American companies like OpenAI, Google, and Meta are at the forefront of AI innovation.

China follows closely, with 11% of top-tier AI researchers and significant private investment, totaling $95 billion between 2022 and 2023 . Chinese companies such as Tencent, Huawei, and Baidu are key players, developing advanced AI models like Tencent's Hunyuan and Huawei's Pangu.

The United Kingdom ranks third, with a robust AI market valued at $21 billion. The UK is home to leading AI startups like DeepMind and Darktrace, and government initiatives are boosting AI research and development with substantial investments in supercomputing and AI infrastructure .

Israel and Canada round out the top five, with Israel's tech scene attracting $11 billion in private investment over a decade and Canada emerging as a significant AI player with $2.57 billion invested in AI research between 2022 and 2023 . Israel is known for its AI-driven cybersecurity and language processing technologies, while Canada hosts innovative companies like Cohere and Scale AI, supported by government initiatives promoting responsible AI development.

This global landscape of AI research and development underscores the collaborative and competitive nature of the field, with each country contributing unique strengths and innovations that drive the technology forward.

AI and Job Displacement

The high percentage of functional illiteracy makes populations more vulnerable to job displacement by AI technologies. As AI systems continue to improve and automate tasks that require basic cognitive functions, individuals who lack strong critical thinking and comprehension skills are at a greater risk of losing their jobs. This displacement is particularly evident in sectors such as manufacturing, retail, and administrative roles where routine tasks are easily automated. Without the ability to engage in higher-order thinking and problem-solving, functionally illiterate individuals may struggle to transition into new roles that require these skills, exacerbating economic inequality and social displacement.

In the United States, where approximately 54% of the population has low literacy skills, the country also leads in AI investment with $249 billion, indicating a high risk of job displacement due to AI advancements . Similarly, China, with 11% functional illiteracy and $95 billion in AI investment, faces significant risks . The United Kingdom, Israel, and Canada also show substantial AI investments ($21 billion, $11 billion, and $8.64 billion respectively) coupled with functional illiteracy rates of 25%, 20%, and 27% . These statistics highlight the potential for AI to penetrate job markets and replace low-skill jobs with automated systems, particularly affecting those with lower literacy levels.

In developing regions, the impact of AI on job displacement is exacerbated by already high rates of functional illiteracy. For instance, in Sub-Saharan Africa and South Asia, where functional illiteracy rates are significantly high, the introduction of AI technologies in sectors like agriculture and manufacturing could lead to widespread job losses. This creates a pressing need for educational reforms and skill development programs to prepare the workforce for an AI-driven future.

The following graph illustrates the relationship between low literacy skills and AI investment across different regions, underscoring the need for targeted interventions to mitigate the potential adverse effects of AI on employment.

Mitigating the Rise of Functional Illiteracy

The intersection of artificial intelligence advancements and the decline in human critical thinking skills poses significant societal challenges. To address the rise of functional illiteracy and its consequent vulnerability to AI-driven job displacement, a multifaceted approach is essential:

1. Educational Reform: Reforming educational systems to prioritize critical thinking, problem-solving, and digital literacy from an early age is crucial. Curricula need to be updated to include not only basic literacy and numeracy but also the skills necessary to thrive in an AI-driven world. Encouraging lifelong learning and continuous skill development can help adults adapt to evolving job market demands. For example, Singapore has launched initiatives to train its population on AI and digital skills, ensuring they are well-prepared for the future job market (The Straits Times).

2. Access to Technology: Bridging the digital divide is fundamental to ensuring equitable access to education and technological resources. Providing affordable and widespread access to digital tools and the internet will enable more people to participate in the digital economy and improve their literacy and critical thinking skills. Programs like the Affordable Connectivity Program (ACP) have been crucial in providing affordable internet access to underserved communities. The ACP offers qualifying households a $30 monthly discount for broadband service, helping 21 million Americans stay connected. However, funding for the ACP is at risk and needs congressional support to continue beyond 2024 (City Limits).

3. Public and Private Sector Collaboration: Governments, educational institutions, and private companies must collaborate to create programs that foster skills development. Public-private partnerships can help fund and develop training programs, internships, and apprenticeships that equip individuals with the skills needed for future job markets. An example is the European Union's Digital Skills and Jobs Coalition, which brings together Member States, companies, and other stakeholders to tackle the lack of digital skills in Europe (European Commission).
   

4. Community Programs: Community-based initiatives can play a pivotal role in addressing functional illiteracy. Libraries, community centers, and non-profit organizations can offer tutoring, literacy classes, and technology training to help individuals improve their skills and become more adaptable to changes brought about by AI. For instance, the New York Public Library offers free technology classes to help bridge the digital divide (NYPL).
   

5. Promoting Critical Thinking: Encouraging a culture of critical thinking and intellectual curiosity can combat the decline in human cognitive skills. Media literacy programs, critical thinking workshops, and public awareness campaigns can help people develop the ability to analyze information, solve complex problems, and make informed decisions. The Stanford History Education Group offers resources to help students critically evaluate online information (Stanford History Education Group).

By addressing functional illiteracy through these measures, society can better prepare for the challenges and opportunities presented by the rise of AI. Ensuring that individuals possess the necessary skills to adapt to technological advancements will not only mitigate job displacement risks but also empower people to contribute meaningfully to an increasingly digital world. This proactive approach will help bridge the gap between AI's potential and human capabilities, fostering a more equitable and resilient future.