This week, of all people, Geoffrey Hinton sort of walked out on artificial intelligence (AI), the very technology he helped develop. He fears that companies like Microsoft and Google aggressively competing to create products based on generative artificial intelligence, the technology that powers popular chatbots like ChatGPT, are racing towards danger. Earlier, too, inventors have distanced themselves from the very technologies they helped create. Alfred Nobel, the inventor of dynamite, and J. Robert Oppenheimer, associated with the atom bomb, were upset with the destructive use of the technologies they helped develop. In each of these cases, their walkouts gave rise to significant movements: the Nobel prize and Nuclear Non-Proliferation Treaty.
Hinton’s reaction is a reminder to the AI industry that they need to take immediate steps to allay growing fears around the industry. AI companies will do well to remember what happened to the nuclear energy industry. The truth is that nuclear power generation has one of the lowest levels of fatalities per unit of energy generated among various sources of it. So far, there have only been three major accidents in the industry, only one of which was catastrophic. But with each accident, the Three Mile Island accident in 1979, Chernobyl disaster in 1986 and the Fukushima disaster in 2011, exaggerated and irrational fears over the industry grew and the emotional distance with end users widened beyond repair. Does the AI industry have the wherewithal to face a possible ‘Chernobyl’ like disaster caused by its technology?
The AI industry can build a bright future for itself not just by mitigating its ill effects, but by building on its positives. It can learn a lesson or two from the automotive industry. All the ills that it creates in the world, from road accidents to carbon emissions, are effectively counteracted by playing up its core benefit: the joy of road transportation.
These days, there is much jubilation about the AI industry, especially since the release of ChatGPT. But this enthusiasm for generative AI is unlikely to last long. This jubilation is only about the ‘creation’ and initial usage of the product. For this glow to last for a long time, generative AI technology has to move to a crucial stage: of adoption, with its products in sustained use. Huge enthusiasm at the ’creation’ stage does not always translate to sustained usage of a new product.
ChatGPT is not a machine whose use is dictated by its calibrations. It is more like a tool which in the hands of a dextrous user can produce usable output. The sustained adoption of this new technology will greatly be determined by the quality of what it generates in response to our requirements. So it is imperative that users of ChatGPT are equipped to use it as effectively as possible.
Individuals who work in a creative industry like advertising know that the quality of the creative output is as good as the quality of the strategic brief provided to the creative team. Soon companies like Google and the Chinese tech giant Baidu are going to launch their own versions of generative AI tools. What will act as a clear differentiator in the generative AI space is the quality of prompts provided to generative AI tools. Those equipped to provide intelligent, creative prompts will get the best returns from these tools.
Advertising agencies usually deploy another strategy to further improve the quality of their creative output: competition. Often, multiple internal creative teams are asked to work on the same strategic brief. This internal competition brings out the best in most people. Taking a cue from this idea, can the output of the neural networks that ChatGPT relies on be used to fire up real neurons in the human brain? Can the output of generative AI, for example, be treated as a benchmark for humans to beat?
Humans have always reacted positively to challenges that new technologies have thrown at them. For example, the most prominent artistic movement in Europe in the early part of the 19th century, as industry arose, was realism.
Representation with the least possible distortion of reality was its focus. Around this time, the technology of photography was also invented, forever changing the nature of visual representation. Photographs could create images that were far more realistic than the best any human artist could paint. So how did human artists and portrait makers deal with this competition from the new technology of photography?.
Painters started began to explore new directions of artistic expression using various dimensions of painting. The artistic movement Impressionism was the first to deviate from the realistic norm. Impressionist painters like Monet’s focused on conveying the essence of a scene through colour, light, movement and emotions. Expressionism, as an art movement, emerged to express the meaning of emotional experiences rather than physical reality. In Cubism, subjects were analysed, broken up and reassembled in an abstract form with multiple perspectives —instead of depicting objects from a single perspective. There was also Surrealistic art, which allowed the unconscious mind to express itself on a canvas.
The emergence of photography as a technology gave rise to so many great artists, like Vincent Van Gogh, Pablo Picasso, Salvador Dalí and many more, who transformed the way we think about art. If so, just imagine the human potential that could be unleashed to take on the power of a technology like generative AI. Geoffrey Hinton would then end up proud of this development.
Biju Dominic is chief evangelist, Fractal Analytics, and chairman, FinalMile Consulting.
Catch all the Business News, Market News, Breaking News Events and Latest News Updates on Live Mint. Download The Mint News App to get Daily Market Updates.