The deflating AI bubble is inevitable — and healthy

The hype and excitement surrounding AI was bound to come back to earth at some point, and this summer was that moment. Those of us old enough to remember saw something similar during the dotcom boom and bust of the late 90s and early 2000s, as well as more recently with crypto.

After rising nearly 200% in the first six months of the year, AI darling Nvidia saw its share price fall 20% during July and August, which tracks with Gartner putting gen AI in the “Trough of Despondency” in its latest Hype Cycle chart. Even Goldman Sachs, previously bullish on the AI story, has raised concerns over whether there’ll be positive ROI for many of the investments being made in the technology. However, these concerns and predictions are inevitable given some of the unrealistic expectations surrounding AI. A little realism and scepticism are healthy at this stage of an emerging technology, but it’d be unwise to assume AI has hit a dead-end.

Beware the naysayers

Predicting the future is generally a fool’s errand as Nobel Prize winning physicist, Niels Bohr recognized when he stated, “Prediction is very difficult, especially about the future.” This was particularly true in the early 1990s as the Web started to take off. Even internet pioneer and ethernet standard co-inventor Robert Metcalfe was doubtful of the internet’s viability when he predicted it had a 12-month future in 1995. Two years later, he literally ate his words at the 1997 WWW Conference when he blended a printed copy of his prediction with water and drank it.

But there comes a point in a new technology when its potential benefits become clear even if the exact shape of its evolution is opaque. We’re at that stage now with AI, and rapidly developing LLMs and their generative capabilities that are steadily diffusing through enterprises. There’ll be further bumps in the road but there are now enough deployments and use cases for us to see the potential for a range of digital transformations across businesses and the public sector.

What are business leaders telling us?

Earlier this year, consulting firm BCG published a survey of 1,400 C-suite executives and more than half expected AI and gen AI to deliver cost savings this year. While most were still at the experimental stage with deployments, around a quarter expected savings in excess of 10%, primarily through productivity gains across operations, customer service, and IT. And while a substantial proportion were concerned about their organization’s lack of progress with AI investments, about half of those put it down to unclear investment priorities, and a lack of available skills and coherent strategies, rather than issues with the AI technologies themselves.

At the latest Davos meeting of economists, policy makers, and business leaders, US KPMG CEO Paul Knopp said, “We’re at a phase where we think generative AI will move from pilots and experiments to implementation and industrialization.” A claim supported at the same event by the CEO of Salesforce AI, which is already seeing productivity gains from its customers’ use of AI tools.

Dispatches from the front line

Financial services firm Klarna publicly stated that 90% of its staff used AI daily with its Klarna AI assistant to handle two thirds of customer service chats in its first month of operation, the equivalent work of 700 full-time agents. Aside from the cost savings, the company claims Klarna AI is more accurate resolving issues, leading to a 25% drop in repeat inquiries and adding $40 million to its bottom line.

Many AI deployments and integrations are not revolutionary, however, but add incremental improvements and value to existing products and services. Graphics and presentation software provider Canva, for example, has integrated Google’s Vertex AI to streamline its video editing offering. Canva users can avoid a number of tedious editing steps to create videos in seconds rather than minutes or hours. And WPP, the global marketing services giant, has integrated Anthropic’s Claude AI service into its internal marketing system, WPP Open. This is used by the company’s 114,000 staff across a network of agencies to help with campaign ideation, content generation, and interpreting complex client briefs.

AI is evolving

As use cases emerge and businesses start to learn what works and where the bottlenecks are, we can expect to see a steady rise in AI deployments. The rise of open-source, smaller models are making customizations more accessible, too. AI repository Hugging Face currently lists almost 135,000 models available under an Apache 2.0 licence, which allows commercial reuse and repurposing. This and other model-sharing platforms offer a vibrant bed of resources for developers to adapt and customize models for specific use cases and industry verticals.

The growing use of retrieval augmented generation (RAG) and knowledge graphs alongside this explosion of open models, as well as APIs from big vendors such as OpenAI and Anthropic, offer enterprises a path to exploit their data resources more safely. IDC’s recent claim that 90% of business data is unstructured presents a golden opportunity for firms to use AI to unlock these untapped assets. To do this securely and efficiently — particularly where mission-critical business processes are involved — will take time, but we can expect major breakthroughs in the next couple of years.

So while there’s definitely a current chill in the air when it comes to AI, those who prepare now will be best placed to succeed when it gets warmer.