There has been "greenwashing" and might there be a risk at some point of a similar issue with how the term "artificial intelligence" is bandied about these days by almost everybody?

Like quite a few journalists assailed by comments about how we’re going to be side-lined by artificial intelligence, I tested ChatGPT the other day. I put in search terms about myself to see what it might write (my narcissism was on full blast). It was mostly accurate, but got a few points badly wrong. I almost cried with relief. ChatGPT can’t come up with new ideas, understand full context, or understand why one story matters but another might be a dead end. I am still gainfully employed.

Artificial intelligence, or AI, is all the rage. It seems to be everywhere. By my reckoning, the AI acronym is catching up fast with “ESG” in its ubiquity in my email inbox. It is “hot.” Whether in Silicon Valley, London, Zurich or Singapore, venture capital money is pouring in (well, it was until recently, given financial market shenanigans).

To give some idea of the buzz, HSBC Global Private Banking and EquBot have launched the Artificial Intelligence Powered Global Opportunities Index. This is a rules-based investment strategy featuring IBM Watson’s AI engine and other patented technologies to turn data into investment insights. In the US, Widom Tree has partnered with NASDAQ and the Consumer Technology Association to develop a bespoke index that identifies and classifies AI-focussed companies, providing a “pure AI investment exposure,” to quote the firm’s website. In early March, Salesforce Ventures, the company’s global investment arm, launched a new $250 million generative AI fund to bolster the startup ecosystem and spark the development of responsible generative AI. This news service has carried research and commentary on how different parts of the wealth sector will benefit, such as in crunching data for KYC checks and to flag potential issues, for example. 

The financial sums are large. The US Department of Commerce has reported that global AI funding reached an estimated $66.8 billion in 2021, doubling its previous figures from 2020.

Wealth managers get through a lot of data to do their jobs. This means new computer-based tools aren't luxuries or science fiction, but hard necessity. For example, the field known as behavioural finance uses a lot of it because of a need to spot patterns and see potential warning signs, such as if a person is showing excessive confidence or undue pessimism, etc. ESG investing, to be viable, needs large amounts of data when screening for carbon "footprints", poor labour standards, questionable boardroom moves, and others. Without the sort of computing power now taken for granted, these ideas cannot fly. 

But there are concerns on what AI really amounts to. When OpenAI launched GPT-4, an upgrade to the ChatGPT platform, it did not meet with universal applause. Over at Bloomberg, for example, columnist Parmy Olsen was scathing: “Artificial intelligence in particular conjures the notion of thinking machines. But no machine can think, and no software is truly intelligent. The phrase alone may be one of the most successful marketing terms of all time.”

Her argument is that when people use the term AI, in nearly all cases what is meant is machine learning, not some sort of “intelligence” as she defined it. All that is going on is that these systems are mirroring huge amounts of text. This is an impressive feat, but it is not intelligence. (It is worth reminding ourselves of the old "Turing Test" named after the famous computer scientist and Enigma codebreaker Alan Turing. Turing proposed that a computer can be said to possess artificial intelligence if it can mimic human responses under specific conditions.)

Beyond the Turing Test, what counts as “intelligence” here? It is more than just being able to remember facts or "sound" like a person having a conversation. Depending on one’s view and philosophy, it touches on qualities such as volitional consciousness (the idea that to think requires effort, a sense of “taking charge” of the path your mind is on), introspection (thinking about the process of how one thinks about something to work out if a process is rational or not), ability to imagine and throw up scenarios, and come up with original ideas, etc. These notions – such as about the existence or not of free will – have divided philosophers since before Aristotle. 

Why does all this matter? It matters – and Ms Olsen made a similar point – because there’s a risk that AI gets oversold and degrades into being a marketing sales term, and causes problems when cynicism kicks in. I see this happening with ESG investing. Also, if people make costly investment mistakes because they entrust their retirement fund to a set of algorithms, it is unclear whether a firm can dodge lawsuits or regulatory frowns because “the machine made me do it". But using the term inappropriately can still lead people astray. That’s not good.

As a fashionable area, AI is drawing in a lot of money. Wealth managers are no strangers to fashion, and as we have seen with the “greenwashing” phenomenon, definitions matter. Framing expectations around what this or that technology can do is important.

For the time being, my prediction is that AI, however defined, will augment humans’ capabilities, but won’t supplant them. That can be wonderful – think of how we don't need to remember things like phone numbers or street maps as much as we did because of the internet. But even here, technology is double-edged. I’d go so far to say that there may be a danger that as our lives become ever more comfortable and we spend so much time on social media and using gadgets, our minds will become flabbier. We need to be uncomfortable, learn new skills and subjects, and develop critical reasoning skills. (How, for example, can a computer develop “healthy scepticism”?)

There are a few worries. A report earlier in March said that a  group of psychologists, two from Northwestern University and the third from the University of Oregon, found via online testing that IQ scores in the US may be dropping for the first time in nearly a century (source: Phys Org, 10 March). A paper was published in the suitably named journal, Intelligence, by Elizabeth Dworak, William Revelle and David Condon. They analysed results of online IQ tests taken by volunteers over the years 2006 to 2018). One could assume that a similar trend might be in place in other countries. It is not entirely clear what is going on. One theory might be that in the late 20th century and early 21st, our lives became easier. Information is easier to find. The sharpness we had to cultivate to learn how to "figure things out" is not as urgent as it was for people, say, 50 years before. Admittedly, this is highly speculative. 

The rapid rise of AI has got some people worried about the impact on society. For example, Elon Musk – Tesla CEO and OpenAI co-founder – has called for a six-month pause in developing systems more powerful than OpenAI's newly launched GPT-4. That's not the first time that the spacefaring entrepreneur and Twitter owner has flagged his concerns. 

Whatever the data says, though, what is clear from the dramas we have seen at Credit Suisse, Silicon Valley Bank or the positive achievements of entrepreneurs and innovators, is that at base they depended on human judgement, and a grasp of reality or evasion of it. A computer cannot "evade" reality, but it also cannot, at least as far as I know, "take full responsibility" for something, either.

For the time being, you are going to have to put up with me.