Is The Fear Of An AI Winter Real Or Imagined?

Play Audio Version

Why is the technology and business media now talking about an imminent ‘AI winter?’ This would not be the first time that AI research has caught the attention of the business community and then suddenly the interest and investment dropped away.

There are some commentators saying that a new AI winter cannot come soon enough. Others are continuing to explore how it can transform their business. Meta recently saw their share price soar because they announced that AI has been used to successfully boost how they sell advertising on social media platforms such as Facebook and Instagram.

Since OpenAI released ChatGPT at the end of 2022 the world has seen a strong interest across all business sectors in how AI – especially Generative AI – could be used to transform existing businesses and to create entirely new services.

ChatGPT has been one of the most popular new AI systems. It still has hundreds of millions of regular users. Since then other major players have emerged, including Meta, Anthropic, and Google.

One of the fears about AI at present is that all these companies are heavily invested in research, but they keep spending more than they are earning, with their eyes on a future prize. Will it ever arrive?

OpenAI is just one example. They will spend over $8bn just this year. Their costs to 2029 are estimated to be around $115bn – over $80bn higher than recent estimates. Meta has been hiring AI researchers on pay packages that look more like sports stars – technology professionals demanding $250m packages.

OpenAI is generating about a billion dollars a month in revenue, but analysts estimate that the company still runs at a loss. From now to 2028, Meta plans to spend over half a trillion dollars on AI research – that’s trillion with a T. Although they are seeing success with ad sales, there is surely no way that they are seeing a return on this investment in the short term.

Many commentators are worried that the research companies are spending big and losing vast amounts searching for the magic tool that will make AI essential. There was a parallel back when the IBM PC was launched. Many managers asked ‘why do I need a computer on my desk?’ Then they saw what a spreadsheet can do and realized that they needed their own personal computer.

Where is the AI version of this experience from the past?

Research from MIT recently found that around 95% of companies testing pilot projects using Generative AI did not scale up the projects – they failed. Almost every AI project failed and was discontinued.

This concern over research costs and the high failure rate of AI projects has worried many executives. Some have started to believe that maybe a business transformation using AI has been over-hyped for the past couple of years.

I believe there are some important points to remember though. Every new technology goes through a period where businesses experiment with it and often can’t find a way to make it work. Either they fail to find a return on the investment or it doesn’t change productivity or create any new business opportunities.

There is now a very strong understanding that most AI project failures are because of managers expecting the systems to work straight from the box – or without a significant effort in reorganizing how data is stored and managed within the business. Focus on your data first and you can expect much better results with AI.

Regarding the concern about investors, there are industry voices arguing that we should worry less about private companies funding the AI arms race. Paulo Carvão, a Senior Fellow at Harvard University, recently noted in an article for Forbes that we should look at the AI world through a different lens today.

Previous AI winters were triggered by the tech companies making wild exaggerated promises, a reliance on government funding for the research, and technology that didn’t work very well outside of the lab environment.

The world is entirely different today, as Mr Carvão said: “The economics of computing have changed. In the past, hardware was too expensive or too limited. Now, costs continue to fall while cloud platforms, specialized chips, and massive datasets are widely accessible.”

AI no longer relies on a small number of government agencies – it is mainly venture capital that funds research today. All these differences, from the funding of research to the hardware itself, means that a collapse in funding and interest in AI is far less likely to happen in the near future.

I think we are now within the disillusionment that often follows a widely-hyped and promoted new technology, but this is not terminal. AI is a transformative technology that will eventually be powering entirely new businesses and industries.

AI can still be a powerful driver of change within many industries, but there is a good lesson for every business leader in what has happened within the customer service environment. In 2023, it was assumed that human customer service agents were no longer needed – AI could handle all customer interactions. Now there is a more general acceptance that AI should be assisting and helping those human agents rather than replacing them. The productivity boost and improved service comes from blending people with technology.

If we accept this, then the predicted AI winter is unlikely to ever arrive.

IBA Group has an experienced practice focused on data management, data analytics, and AI. Check the website here for case studies and examples of projects that have been designed and delivered in addition to ideas, insights, and project suggestions.

Follow IBA Group on LinkedIn for regular updates and comment. For more information on technology strategy and how tech connects to real business solutions please click here