Experts who, just two years ago, were quick to hail large language model-based AI as the next big revolution are now suddenly writing it off. They are wrong. AI is here to stay…and disrupt.
If you want to figure out AI’s true potential, then you need to look beyond the companies building the models. (Pic: Reuters)
Eliezer Yudkowsky is a young (44) but influential AI researcher and founder of the Machine Learning Institute in California. His research focuses on AI safety. And one of the things he is fond of saying is this: “By far, the greatest danger of artificial intelligence is that people conclude too early that they understand it.”
Yudkowsky alludes, of course, to the grave danger of not paying attention to everything that could go wrong with AI’s indiscriminate use. But now as AI experts appear to suddenly grow skeptical of the technology’s ability to create value, or even wealth, it is worth viewing Yudkowsky’s quote from a different perspective. The people ready to write off AI (The Economist newspaper appears to do so in one of its recent issues, mainly by pitting valuations of AI companies against their AI income) as just a fancy tool are perhaps missing a simple fact: that, just like many previous inventions—be it the motor car or the airplane or computers—the value of the invention often does not lie in itself. Rather, its value lies in its application.
An airplane parked in the airport hangar is a money loser compared to the one that is flying with passengers on board. An idle computer does nothing, but in the hands of an engineer, scientist, artist, or even a humble fisherman, it produces something of value. The point is, if you want to figure out AI’s true potential, then you need to look beyond the companies building the models…the OpenAIs, Anthropics, or even the Googles and Amazons. You need to look at AI users. Businesses and people who will be using it to do things better, faster, more efficiently, and in entirely new ways.
On that count, it’s early days yet for AI. However, even the initial results do suggest that humanity will do more—and not less—with AI. In every conceivable field. From space research to life sciences to finance to agriculture to art. For that to happen, AI models will need to get better. And they will.
The only question then is, will everyone who has jumped onto the AI bandwagon make money? Of course not. In every industry, the best companies end up with the largest market share and, hence, the biggest share of the value the industry creates. The companies that don’t have the best to offer will capture significantly less value, and many others, none at all. These are the ones that will shut shop.
Therefore, the risk in the AI business is not of value creation. The user ecosystem will figure out ways to profit from it. On the contrary, the risk is that this narrative (of AI being an endless money-guzzler for its developers) could remove or weaken the much-needed guardrails for AI. And that is the only price we as humans must not be willing to pay.
Follow us on social media