Stay informed with free updates
Simply sign up to the Artificial intelligence myFT Digest — delivered directly to your inbox.
The data company Gartner is famous for mapping the “hype cycle” of different technologies. In this year’s edition, generative artificial intelligence has passed over the Peak of Inflated Expectations and is now sliding into the Trough of Disillusionment. Only later will it reach the Slope of Enlightenment and the Plateau of Productivity.
The launch of OpenAI’s ChatGPT three years ago certainly triggered an avalanche of excitement about the possibilities of generative AI. The take-up of the technology has been among the fastest in history. ChatGPT now has more than 800mn weekly active users, according to the company. Users have marvelled at the chatbot’s uncanny ability to perform tasks as varied as writing plausible sonnets about your pet goldfish, summarising complex legal documents or generating passable corporate presentations.
But these foundation models also exhibit some glaring flaws, most notably their tendency to hallucinate or, more accurately, confabulate facts. On countless earnings calls, corporate bosses have extolled the possibilities of deploying AI across almost every business function to improve productivity. But they are also wary of the risks generative AI can pose to data security, client confidentiality and corporate reputation. The excitement aroused by the deployment of AI agents has also run into the hard wall of reality, where nothing is as simple as coders imagine.
Several recent reports suggest that generative AI has so far failed to deliver on early sky-high expectations and corporate disillusion is spreading.
When prompted to identify the “fails” in large language models, ChatGPT itself responds: “The technology itself is potent but without the right preparation it becomes a liability rather than an asset.” That seems a pretty good summary of where we stand today.
Yet with the right preparation and sensible deployment, generative AI can be an impressive productivity tool.
This is perhaps most apparent in the performance of some of the big US tech companies that best know the possibilities and flaws of the AI models. As research company Alpine Macro has pointed out, these companies are currently enjoying a “jobless profit boom”, reflecting accelerating productivity growth.
The tech companies may have overhired during the Covid pandemic when the world went online, but since then they have been shedding jobs. Alpine Macro notes, however, that technology-related jobs have been in recession for more than three years, suggesting something else is behind it now. “We suspect that job losses in tech have been driven mainly by AI displacement,” Chen Zhao, the company’s chief global strategist, writes.
Intriguingly, this phenomenon appears to be spilling over into the broader economy. Private sector employment in the US remains 5 per cent below its pre-pandemic trend, according to Alpine Macro.
The mass deportation of undocumented workers, reducing the labour supply, may also have prompted more companies to invest in tech. Productivity growth is now more than twice as fast as it was in the 2010s, Alpine Macro estimates.
So how can the individual company maximise the potential benefits of AI while minimising the risks?
One business that has been determinedly applying AI is Mimecast, a global cyber security company with more than 40,000 customers. The company has been using AI both to enhance its service offering, helping to detect cyber threats, and to boost its own corporate productivity.
All of Mimecast’s 2,400 employees have been encouraged to adopt the technology and extensive training has been provided to ensure its ethical and responsible use. Mimecast has been working with training start-up Pair to assess its employees’ capabilities and to upgrade their skills.
Tim Seamans, Mimecast vice-president for AI and business transformation, says that the company’s workforce could be divided into early adopters, the cautiously optimistic and the sceptics.
But the company has systematically helped move its employees up the AI value chain. Now 96 per cent of its staff have incorporated the technology into their daily workflow, significantly improving productivity. “It’s in the hands of everybody,” he says.
Change has been persistently driven from the top of the company, with the chief executive creating his own AI agents and sharing them with colleagues to prompt discussion about business use cases.
Mimecast has also benchmarked every department against industry adoption rates, exposing the company’s strengths and weaknesses. Often the biggest gains can come in less obvious areas, such as customer service, finance, human resources and sales.
As Seamans notes, and many other executives emphasise, adopting a powerful new technology is not just about figuring out how to use the technology itself. It also involves changes in corporate culture, work practices and business organisation. Only then can companies ascend the Slope of Enlightenment.
