Generative AI has been predicted to add trillions to the world economy in a productivity boost never before seen in history (if it doesn’t wipe out humanity first).
But what if it doesn’t?
A growing number of sceptics, including some leading AI scientists, are wondering whether the tech might not deliver on its promises to boost the world economy.
Goldman Sachs famously predicted that generative AI would bring about ‘sweeping changes’ to the world economy, driving a $7 trillion increase in global GDP and lifting productivity growth by 1.5 percent this decade.
Professor Gary Marcus of New York University wrote on Substack that ‘we are starting to see signs’ that generative AI might be a ‘dud’.
Among the warning signs was a report in the Wall Street Journal suggesting that customers found the $30 a month price of Microsoft‘s new AI-boosted Copilot software too expensive.
Marcus wrote on his Substack: ‘Putting Large Language Models [like ChatGPT] into production is hard; most work to date has been preliminary.
‘Companies are starting to temper their expectations. Many initial expectations were unrealistic.’
Marcus points out that progress towards actually making money out of Large Language Models has been slow – and that after OpenAI launched GPT-4 in 2023, no one has launched a model that is decisively more powerful.
We might be reaching a plateau in terms of sheer capability. Nobody has been able to beat it decisively. Places like Google and Anthropic have put in a lot of money trying. None succeeded; instead there tentatively seems to be convergence at GPT-4 levels.’
Marcus says there has been ‘progress’ towards finding possible uses for Large Language Models – but so far the technology is often being used for crime.
He said: ‘Bad actors, who may have lower standards for reliability, appear to be using them for cybercrime and disinformation.’
Retired professor Jeffrey Funk points out that AI has spent huge amounts of money trying to combat ‘hallucinations’ – where AI systems ‘make up’ facts – but has not solved the problem.
He wrote on LinkedIn: ‘The revenue isn’t there yet, and might never come. The valuations anticipate trillion dollar markets, but the actual current revenues from generative AI are rumored to be in the hundreds of millions. Those revenues genuinely could grow by 1000x, but that’s mighty speculative. We shouldn’t simply assume it.
Funk also warns that the pace of innovation in generative AI appears to be slowing.
He said, ‘Think about PCs or the iPhone. There were big improvements in system performance during the early years that diminished over time despite 40 percent annual improvements in the performance to price ratios of memory and processor chips.
‘With Moore’s Law having slowed considerably in the last 5 to 10 years, [OpenAI’s Sam] Altman can’t expect much more from Moore’s Law and those concerned with generative AI’s big appetite for energy are bound to push for regulations.
Speaking to The Information, Todd Lohr, a principal at consulting firm KPMG, which resells Microsoft products, was lukewarm on the benefits of Microsoft’s CoPilot AI products.
Lohr said: ‘Word is okay, Powerpoint isn’t particularly useful unless you train it on specific [instructions] because it only creates a Powerpoint that’s very basic.
‘Excel is not there yet—you have to spend a ton of time prompt-engineering to get it to do anything for you, which takes way longer than just writing the Excel formulas yourself.’
Bank of America investment strategist Michael Hartnett has previously suggested that AI might be a bubble, comparing it to the dotcom crash of 2000.
Amazon CEO Andy Jassy said during an earnings call in February that near-term revenue from AI is ‘relatively small’.
Speaking to DailyMail.com, Dom Couldwell, Head of Field Engineering at DataStax, said that we remain in the ‘unknown unknowns’ phase of generative AI.
Couldwell said: ‘This area has seen so much hype, it is growing up in public – it took Netflix three years to get to one million users, but it only took ChatGPT five days.
‘There are also companies looking at this as the next get rich quick scheme after crypto.’
The companies that Coudwell works with are still trying to find where Generative AI can deliver results, he said.
He said: ‘Not to get too technical, but the challenge is how companies get their own data and Intellectual Property working for them, rather than just leveraging OpenAI or Google’s tech.
‘Just trying to get value from GenAI from out of the box solutions or to replace staff won’t work – you need to use it as a multiplying factor to make your employees more productive, offer more value to customers and differentiate from your competition’s cut and paste chatbot.’