I first encountered artificial intelligence (AI) at the Pearl Market in Beijing. The market at the seat of Chinese Communism is one of the most extreme examples of free-wheeling capitalism you’ll find anywhere, with six floors of stalls selling everything fake.
Half the joy is haggling. The Chinese sellers start high and the game is to bargain them. It’s pure theatre, full of dramatic smears, fake outrage, grimaces, winks, nods and finally deals. The place is busy.
The saleswomen are Chinese and speak amazing English. I went with a friend from Argentina who claimed not only to be fluent in Spanish but also to know Latin American idioms. We couldn’t believe it. They chatted effortlessly about the deal, the quality, the origin of the fabric, the brands, the price differentials, the factories, prices in New York, London and yes, Dublin. They giggled and played in perfect English. Likewise in Spanish.
It was an extraordinary experience until we asked them questions like, “What’s the weather like in Beijing in spring?” Out of complete fluency, when they discussed prices, brands, exchange rates and percentages, they looked at each other blankly, then at us .
It later struck me that these impressive saleswomen were somewhat similar to human “language bots” who had skillfully learned some English and Spanish (plus French and German) and were able to appear fluent, but could only talk about a narrow range of topics – in this case, mainly the sale of jeans. They had been taught to anticipate questions from foreign buyers based on the likelihood that we would all be asking the same questions.
That was their dataset, if you will, and they had learned to recognize the sounds of a language they didn’t understand, rather than actually understand it. It turned out that there was a language school on the top floor of the market that many of them had attended, at least that’s what I was told.
The intelligence of this robot is breathtaking. Ask him to write an essay, article, or blog, and the quality of the material he spits out is impressive — and frightening
It was still phenomenal to witness and most people marveled at the linguistic dexterity of these imitation brand salesmen.
Since then, the explosion of e-commerce and data-mining has made “bots”—machines that are triggered by prompts and anticipate what questions a human will ask—a part of our online lives. Her competence grows day by day.
For example, this week I tried the ChatGPT app. The intelligence of this robot is breathtaking. Ask him to write an essay, article, or blog, and the quality of the material he spits out is impressive — and frightening. The robot scours the internet for material on the subject and then, as if it were a learned human brain, structures all this information into a coherent ‘thinking process’ that produces the desired number of words in a matter of seconds.
AI will be part of our future, impacting certain professions the way Gutenberg’s printing press impacted the legions of monastic scribes who wrote, illuminated, and published manuscripts before the printing press. That is, while scribes may have been outmaneuvered by the printing press, the same technology led to an explosion of books, education, and ideas. The world benefited while the monks suffered. AI will be the same. It’s already there. Look at these numbers.
According to Statista, the revenue from the AI software market is expected to reach €118 billion ($126 billion) by 2025. Three quarters of all customer service interactions are expected to be driven by AI-based platforms by 2025. According to a survey conducted by Deloitte, the proportion of companies that say they use three or more types of AI has increased from 62 percent to 79 percent in just one year. Nine out of ten business leaders agree that AI will be critical to success in the next five years. According to a PWC report, AI could contribute as much as $15.7 trillion to the global economy by 2030
The application I’ve been using this week, ChatGPT, was launched by OpenAI in November 2022. It’s looking for more capital to expand, expecting $200 million in revenue next year and $1 billion in revenue by 2024. Bloomberg has reported that Microsoft plans to pump $10 billion into the company.
Excited investors are no guarantee of success – remember the hype surrounding the Metaverse this time last year? But the uses for AI are many, and some may surprise you. For example, earlier this week, on Valentine’s Day, there’s a good chance that the love letter you received was written by a robot. If that wasn’t the case this year, it will be next year. According to this 2023 survey conducted by McAfee, one in three men and about one in five women worldwide planned to use AI to write love letters this Valentine’s Day, with 21 percent of respondents saying, “I’m short on time and that would be most efficient,” she explains. Only 30 percent said they could tell the difference between a robot-generated love letter and the real love letter from the human heart.
Looking to the future, this is where one of the many clashes between AI and humans will take place. What happens to society if we outsource subjects like history, languages and literature to robots in the education system and allow them to scour the internet for information, and we encourage students to explore technology, science and math?
While AI robots are capable of reproducing information that is already known, they are still unable to create emotional connections based on human experiences
Every scientific breakthrough has had the effect of moving people from using deductive reasoning, which relied on what is already known, to inductive reasoning, which projects forward rather than backward. AI will likely push us further towards inductive reasoning.
Fortunately, while AI robots are capable of reproducing information that is already known, they are still unable to make emotional connections based on human experience. New ideas come from critical thinking and people make complicated and unexpected combinations. That is still the domain of humanity. But how long?
One thing is clear: the development of AI and the like from ChatGPT means we need to move our education system away from memorization and burping. This is a major challenge to Leaving Cert’s approach to education. A test system that is largely a memory test has to go. It needs to be replaced with something less rigid and ultimately less exam-based. The Leaving Cert must at least attempt to prepare students for real life. In a world of AI, the Leaving Cert is redundant.
It’s time to embrace 21st century thinking. If we don’t do it, the robot will.
