Artificial intelligence has been sold to the public as a revolution on the scale of electricity or the internet — a technology that will replace workers, transform entire industries, and fundamentally alter how society functions. But is AI actually living up to the hype?
The only tangible thing artificial intelligence has actually done for most businesses — since ChatGPT hit the market in late 2022 — is increase efficiency for some tasks, such as summarizing documents, generating drafts, automating customer service responses, and helping some programmers write code faster. In some cases, however, the oversight and editing required to ensure accuracy of those tasks takes longer than completing the work from scratch.
AI still hallucinates facts, makes up quotations when writing news articles and legal documents (even when instructed not to), and delivers prose that, at its best, resembles that of a strong high school freshman. Yet AI has been framed as a substitute for expertise rather than a tool to supplement it, and the prevailing message emerging from Silicon Valley’s never-ending AI symposium is that artificial intelligence is expected to replace journalists, teachers, programmers, artists, and customer support workers — in short, nearly every profession except plumbing.
If anything, AI is a tool, and like most tools, serves as an extension and/or amplification of human capabilities. For instance, a wheel is an amplification of the foot, clothing is an extension of the skin, the internet is akin to a brain, and AI is an extension of human consciousness. And while AI may be an extension/amplification of human consciousness, it does not supplant it.
At the present time, AI is certainly not better than a seasoned professional with at least 10 years experience in any field. For AI to outperform someone at their job, that person would have to lack basic skill, judgment, competence, and/or be a complete stiff. Fully automating my newspaper with the AI tools that are presently on the market, for example, would put it out of business. There is no desirable job in the near future that AI is realistically positioned to eliminate. In fact, some of the biggest recent layoffs have hit AI workers at AI-focused tech companies.
And while billions of dollars have been dumped into many of the largest AI developers — including OpenAI, Google DeepMind, and Anthropic — those companies are blowing through much of that money and more to train and operate AI models that are expensive to run and difficult to monetize sustainably. The Harvard Business Review has noted that many AI companies still lack profitable business models, and as the cost of operating large-scale AI server farms continues to climb, Jeff Bezos has predicted that some energy-intensive computing infrastructure could move into space within the next decade as a way to address the growing resource strain of AI computing.
To date, the public narrative around AI is still largely promotional in nature. The technology is real, and it is useful, but what it delivers today is not what it has been marketed to deliver.