Lessons from past bubbles for the AI revolution

THE artificial intelligence (AI) revolution, with its expansion into neural networks and other novel fields, marks a dramatic shift away from traditional innovation models.

Like all revolutions, it comes with challenges, as rapid technological advancement gives rise to concurrent risks. Market volatility and convoluted regulations are significant hurdles, especially for generative AI and large language models.

But previous market bubbles provide valuable lessons for investors and emphasise the need for a clear-sighted, cautious approach.

Setting the pace

Today’s AI trends are influencing both the macroeconomic outlook and our investment strategies. With their enormous influence, Google, Microsoft, Meta, IBM, Amazon, Nvidia and other technology giants are setting the pace for the rapidly evolving sector.

By nurturing specialised AI startups and continuously innovating and delivering new AI products, these companies are laying the foundation for the industry’s future.

While progress is substantial, especially in graphic processing units, the slow pace of mass adoption is a concern. By deploying open AI models, however, Big Tech could help bring stability to the market. AI has had a relatively small direct impact on Big Tech’s revenues, but contributed a projected US$2.4 trillion increase to the sector’s overall value.

Generative AI has an undeniable appeal. ChatGPT and other platforms have made remarkable strides with their conversational prowess. Yet, they betray a surprising lack of depth. They build sentences based on statistical patterns, not deep comprehension. Such a flaw could contribute to the spread of misinformation.

Need for guardrails

Despite such shortcomings, investment capital continues to flood into these systems, propelled as much by AI’s buzzword appeal as its evidence-based results. The disparity between public perception and practical utility is marked, but generative AI is poised to up its game in the years ahead and address its limitations.

Few sectors are immune to generative AI’s potential benefits. As the technology is honed and deployed at scale for commercial use, the productivity gains across the global economy could be astronomical.

While generative AI is shaping market trends, significant regulatory impediments are coming into focus, particularly around the transparency of algorithms, and underscoring the inherent risks. That’s why AI investors should be on the lookout for companies with solid fundamentals and pragmatic valuations as a hedge against the uncertainties embedded in the market.

As AI investors, we must be discerning. Not all AI startups are sound investments. For example, Lede AI’s venture into AI-generated news articles was a disappointment. AI-generated journalism missed critical details, injected inaccuracies into its stories, damaged the reputations of storied news organisations, and underscored AI’s quality and consistency issue.

Online education company iTutorGroup applied AI to its recruitment processes and subsequently had to settle an age-discrimination lawsuit, emphasising why AI applications require robust guardrails to avoid such financial and reputational traps.

Reality is creeping into the AI sector in the wake of the ChatGPT boom. Jasper and other emerging companies have grappled with dwindling user engagement and workforce cutbacks. Platforms such as Midjourney and Synthesia have seen diminished traffic as they dialled back their ambitions for market dominance.

Now, many AI applications would be satisfied with proficient functionality. The strong positions of tech giants, including Microsoft and Google, have also given investors pause.

A stark gap has emerged between high-flying investor aspirations and genuine market conditions. The enthusiasm that spurred the initial wave of AI commercialisation is giving way to disillusionment and doubt.

The high cost of AI model training and the lack of a transparent and viable business blueprint have contributed to the growing frustration, as have a host of legal and ethical debates. Given such difficulties and despite a significant influx of capital and widespread public anticipation, AI startups may be hazardous investments.

Shift in control

US President Joe Biden’s Oct 31, 2023, executive order signals an imperative shift in the control of generative AI. It seeks to position the US at the forefront of AI development and emphasises safety, security and addressing algorithmic bias.

The order requires AI developers to conduct safety tests and publicly share their findings. It holds the US Department of Commerce and other entities accountable for defining and regulating AI standards.

While these mandates will help ensure AI’s safe and ethical application, they could also further increase execution costs, slow research and development, and impose new standards on data privacy and management.

Such regulation could limit AI’s application, particularly among smaller firms and startups, potentially stunting their growth. Finding the right balance between AI development and the essential supervisory role of public policy will be an ongoing challenge for US and global regulators.

Beware the bubble

In today’s high-speed, tech-driven investment world, bubbles are both more frequent and more intense. The main accelerant is the pervasive influence of the Internet and social media. This dynamic ensures the rapid flow of capital into developing trends and fuels the cyclical fervour of AI investment.

What are the implications of this? A likely procession of booms and busts within the AI sector resembling generational shifts, with each surge and downturn shaping and propelling the industry’s evolution.

Does this mean investors ought to pull back? Certainly not. Rather, it underscores how crucial an intelligent investment strategy in emerging AI technology could be. We must exercise thorough due diligence and keep a keen eye on cash flow and other solid value indicators. Exposure to investments rooted in unrealised and unproven potential should be carefully controlled.

Technology bubbles are nothing new. From railway mania in the UK to the dot-com bubble in the US, they underscore the interplay between economic theory and speculative fervour. Bubbles can end in swift, dramatic market implosions or gradual deflations, and they can transform entire industries. Despite the excessive speculation, many present-day tech leviathans emerged out of the dot-com bubble and went on to reshape our world.

The dot-com boom reminds us of the dangers of unchecked optimism when investing in technology. But we must also remember the tech industry adapted and refocused on the intrinsic value of its investments. This period of fine-tuning underscored the industry’s resilience and versatility.

After all, despite consistent growth and industry dominance, Microsoft and Amazon haven’t been immune to the boom-and-bust cycle.

Between 1990 and 1999, Microsoft’s shares surged 10,000 per cent, from US$0.60 to US$60, only to plunge 60 per cent as the dot-com bubble burst. It took years before the company clawed its way back to its 1999 market valuation after bottoming out in 2009.

Amazon’s stock, meanwhile, fell more than 90 per cent amid the dot-com crash and didn’t revisit its 1999 high until 2010.

So, while we may be tempted to ride the wave of skyrocketing tech stocks, we need to temper our enthusiasm with caution and sound judgment. Tech bubbles are unpredictable and potentially destructive. They transform industries, propel substantial progress, inspire much-needed policy reforms, and promote vigilant investment practices.

They have been essential to human progress. But very few tech ventures last, even if they serve as stepping stones to further innovation.

The ebbs and flows of generative AI growth do not necessarily signal severe market instability; they are inherent characteristics of technological evolution within a market economy.

The rise and fall of the fibre optic and 3D printing industries demonstrate how these phases catalyse future advancements. Despite their volatility, electric vehicles, renewable energy and other sectors have developed, driving down costs and leading to widespread adoption.

We have to keep this in mind and approach AI development with a sense of equilibrium. This will help us rein in the risks as we invest in AI’s vast potential and pave the way for a future where technology evolves within ethical and sustainable parameters.

The writer, CFA, is a financial professional with a background in banking and energy