Email Icon - Elements Webflow Library - BRIX Templates

Subscribe to our weekly email newsletter!

Check - Elements Webflow Library - BRIX Templates
Thanks for joining our newsletter
Oops! Something went wrong while submitting the form.

Artificial Intelligence: The Great Lie

Austin Meusch
November 9, 2023

The chatter about chatbots like ChatGPT has reached a fever pitch. Some hail them as marking a new era of transformative artificial intelligence. But how much of this enthusiasm is overhyped marketing? In reality, ChatGPT represents an incremental evolution in language processing, not the dawn of thinking machines. This article will cut through the hype and explore these systems' capabilities and limitations in clear terms.

Silicon Valley’s Big Ego

For decades, Silicon Valley has been the epicenter of transformative technology and business innovation. Companies like Apple, Google, and Facebook were born there and grew to become tech titans. But in recent years, some of these giants have faced reputational challenges over issues like privacy, misinformation, and antitrust concerns. 

As Silicon Valley's halo has dimmed, it has hungered for its next big breakthrough to restore its image as the world's premier tech innovation hub. Enter ChatGPT in late 2022. Silicon Valley quickly crowned it the latest and greatest artificial intelligence. But this label seems more ego-driven than reality-based.

ChatGPT is powered by a large language model, not true artificial intelligence. It cannot reason or think originally - it simply generates responses based on its training data. While impressive, it has major limitations. But Silicon Valley inflated its capabilities to position it as a revolutionary AI achievement.

Some tech leaders used hype and fear-mongering, warning that ChatGPT could either profoundly improve society or dangerously end it. This melodramatic narrative served the ego-driven goal of restoring Silicon Valley's reputation as the scene of humanity-changing innovations. 

In truth, ChatGPT is an incremental evolution not a sudden revolution. Responsible development is needed, not hyperbolic warnings. Silicon Valley's ego led it to exaggerate ChatGPT's importance. Large language models have real potential but true AI remains science fiction, not current reality.

Chat GPT: What it is and what it is not

ChatGPT is powered by a large language model (LLM) - essentially a massive database of text extracted from the internet and books, used to train a complex statistical model. This model can generate human-like text by predicting the next word in a sequence based on the patterns in its training data.

Some have mistaken ChatGPT's eloquent responses as evidence of true artificial intelligence. But while impressive, ChatGPT does not actually think, understand, or reason like humans do. It lacks a model of the world, consciousness, emotions, or common sense.

ChatGPT excels at language processing and text generation within the scope of its training data. But it has no real intelligence or agency of its own. It is an advanced auto-complete program, not an artificial general intelligence.

Because so much modern communication occurs via text, it is easy to anthropomorphize ChatGPT and imbue it with human-like cognition. But at its core, it remains a database of text statistics, not a thinking entity.

The most effective users of ChatGPT treat it as a tool, not an artificial human. They understand its limitations and thoughtfully design prompts and use cases focused on its strengths in language processing. Expecting human-level thinking from ChatGPT leads to disappointment. But utilized properly as a text-based calculator, its capabilities are powerful.

Moving forward, we must appreciate what large language models like GPT are, and are not. Impressive text generation, yes. True intelligence, no. Hard problems remain before we achieve artificial general intelligence. ChatGPT is an incremental step, not a giant leap.

Fear and Regulatory Capture

Some prominent AI leaders like Sam Altman and Elon Musk have sounded dire warnings about the dangers of artificial intelligence in repeated media interviews. They argue that unchecked AI poses an existential threat to humanity, and that strong government regulation is needed now before it is too late.

But are these urgent calls for AI regulation really motivated by concerns for public safety? Or do they represent an attempt at regulatory capture - when business interests lobby for regulation that ends up entrenching their dominant market positions?

There are legitimate risks with rapidly advancing technology like AI that merit thoughtful governance. But this morbid rhetoric and demands for sweeping new rules from some AI executives seem intended less to protect the public, and more to benefit large incumbent AI firms.

Excessive regulation could make it harder for new startups and researchers to innovate in AI. And rigid rules defined by today's limited understanding could constrain promising applications in the future.

Rather than reactive rule-making driven by fears of worst-case scenarios, the public is better served by proportionate governance that balances safeguards and innovation. And we should view appeals for regulation from AI industry leaders with skepticism about their motives.

True leadership in AI safety comes from enabling broad research and democratizing access, not imposing top-down constraints to benefit large incumbents. The public interest requires vigilance to distinguish responsible governance from self-interested regulatory capture.

Large Language Model vs. Artificial General Intelligence

A large language model like ChatGPT is a sophisticated statistical system trained on massive amounts of text data. It can generate human-like writing by predicting probable sequences of words based on the patterns in its training corpus.

But despite its linguistic prowess, a large language model lacks true intelligence or understanding about the world. It has no integrated conceptual knowledge, common sense, reasoning ability, or grasp of causality. It simply predicts next words statistically.

In contrast, the hypothetical goal of artificial general intelligence (AGI) is to create an AI system with the breadth of intellectual abilities that humans possess. This includes not just language facility, but integrated knowledge, reasoning, planning, social skills, general problem-solving abilities, and more.

Current large language models may sometimes give the illusion of human intelligence when interacting conversationally. But their capabilities are narrow, brittle, and unreliable outside their training distribution. True AGI remains firmly in the realm of theoretical speculation for now.

As impressive as systems like ChatGPT are, conflating them with human-level AI is misguided. They operate via pattern recognition on textual data, not integrated reasoning. For now, large language models are useful tools. But the grand challenge of developing artificial general intelligence remains unsolved. We should be clear-eyed about the difference.

A Practical Use Case: Language Education

While hype around large language models like GPT runs high, finding beneficial real-world applications takes clear eyes. One promising use case is leveraging their vast text exposure for language education.

LLMs can provide customized feedback on grammar, vocabulary and sentence structure for students learning a new language. Their pattern recognition strengths are well-suited for translating texts between languages and identifying errors. This application as an education tool shows promise.

However, some early proposed use cases like customer service chatbots have proven ineffective so far. Unlike humans, LLMs cannot truly understand customers' problems or care about resolving them. They simply generate statistically likely responses, not solutions.

As we experiment with applying large language models, we must recognize their limitations. They lack human context, reasoning and empathy. The most valuable uses will be specialized applications in fields like education that play to their pattern recognition strengths.

But we should reject uses that anthropomorphize LLMs as caring assistants. At the end of the day, these are machines driven by data statistics, not hearts. The onus remains on humans to direct LLMs towards beneficial purposes that augment our capabilities rather than replace our humanity.


The arrival of large language models like ChatGPT is an engineering marvel worthy of attention. But inflated claims about creating artificial general intelligence do more harm than good. These systems have real strengths in specialized applications like language education, not in replicating human cognition. True AI remains beyond reach for now. Rather than fear or hype, the wise path is pragmatic experimentation. If guided by ethical priorities and clear eyes, not profit motives and marketing buzz, large language models can empower people and enhance knowledge - while remaining useful tools, not artificial brains.

Create your free Account today to get started

Share on Twitter Share on Facebook Share on LinkedIn