Digital Economy Dispatch #172 -- Living in an AI Bubble: A Survival Guide

Digital Economy Dispatch #172 -- Living in an AI Bubble: A Survival Guide
25th February 2024

Another day, another set of headlines proclaiming the disruptive effects of new digital technology, questioning the future of society in world in digital flux, and championing the power of the latest AI systems and tools to change all our lives. The hype around AI is reaching fever pitch.

The last few years have witnessed a meteoric rise in the development and adoption of AI. And inevitably, this has been equalled by the amount of hype surrounding it. From self-driving cars to automated customer service bots, AI has captured the imagination of the public and industry leaders alike. However, amidst this frenzy, concerns about an AI bubble loom large. Similarities with the “dot com bubble” of the 1990’s are being highlighted. There are even comparisons being made with the Dutch tulip bulb bubble in the 1600s.

No surprise, then, that this is prompting questions about inflated expectations and the reality of achieving the potential of this transformative technology. As digital leaders navigating this dynamic landscape, we must balance these concerns by adopting a critical yet optimistic lens to understand AI's true value proposition and overcoming its challenges effectively.

Are we experiencing an AI bubble? Undoubtedly. But perhaps what matters now is learning how to make the most of this time of change, and planning to exploit what gets left behind once the bubble pops.

Hype vs. Reality: Deconstructing the AI Bubble

Let's be clear: AI is not a magic wand. While its potential is undeniable, the current narrative too often paints an unrealistic picture of its capabilities. Exaggerated claims about superhuman intelligence and imminent robot takeovers fuel a dangerous hype machine. This hype, coupled with inflated expectations, is leading to disappointment and disillusionment when the reality of AI's limitations and current state of development becomes apparent.

Such challenges are most clearly seen when viewed from the perspective of large-scale adoption of AI in real-world scenarios, particularly in legacy-laden public and private sectors. Here, significant issues must be overcome, including:

• Data Challenges: Access to clean, reliable, and properly labelled data is crucial for effective AI implementation. Legacy systems often lack this infrastructure, hindering the development and deployment of robust AI solutions.

• Talent and Expertise: Building and maintaining AI systems requires specialized skills and expertise in data science, machine learning, and domain knowledge. The talent gap in these areas presents a significant hurdle.

• Integration and Interoperability: Integrating AI solutions with existing systems and workflows can be complex and time-consuming, especially in large organizations with diverse IT infrastructures.

• Ethical Considerations: Concerns about bias, fairness, and transparency in AI algorithms are paramount. Addressing these concerns through responsible development and deployment practices is crucial.

• Regulatory Uncertainty: The evolving regulatory landscape surrounding AI adds another layer of complexity for businesses seeking to adopt this technology.

All of these issues, and more, reveal significant concerns about delivering digital change of any kind. With the escalating expectations surrounding AI, the tensions become much more intense as the gap between aspirations and reality grow.

A Focus on What's Left Behind: The Positive Ripple Effects of AI

In an insightful article written by Cory Doctorow at the end of 2023, he directly addressed the issue of what it means to be experiencing an AI bubble. In addition to emphasizing the negative aspects that this may bring, Doctorow argues that the aftermath of such bubbles can yield significant value. For example, he explains how WorldCom's fibre-optic bubble left us with important capacity and usable infrastructure. He urges leaders and decision makers to look at the AI bubble in the same light.

From his perspective, the AI bubble, characterized by massive investment, has led to a temporary surge in user satisfaction, fostering playful communities around AI tools. However, he questions the sustainability of this model, highlighting the exorbitant costs associated with creating and maintaining large AI models. He raises concerns about the financial viability of currently proposed uses of AI. AI’s main business model is focused on productivity gains and cost reduction It faces challenges when the current AI usage model demands expensive human oversight due to AI's tendency to generate inaccurate results.

Consequently, he predicts a shrinking market for AI applications, particularly in high-profile safety critical projects where costlier AI (requiring substantial human oversight) may not find widespread acceptance. Similarly, he sees equal challenges succeeding with low value, high-cost applications, expressing scepticism about their viability. Furthermore, increasing AI regulation will necessarily reduce AI’s immediate industry impact.

He urges us to look beyond the AI bubble. He sees the survival of smaller AI models like Hugging Face and Llama, operating on commodity hardware. The potential for federated learning and the proliferation of individuals skilled in statistical analysis and AI frameworks like PyTorch and TensorFlow. All of these are seen as positive outcomes.

In particular, Doctorow calls attention to the lack of discussion amongst digital leaders about how to ride this wave to gain maximum value from the AI bubble. He questions the sustainability of current AI business models and advises cautious optimism about potential positive outcomes in terms of smaller models, federated learning, and skill development in statistical analysis.

While Doctorow’s warnings and advice are important, from my own personal perspective there are 3 broader considerations we all should recognize as essential to survive and thrive in the current AI bubble. Despite the challenges, the AI hype is continuing to have several positive consequences:

1. Acceleration of Digital Technology Innovation: The hype surrounding AI has spurred significant investment in research and development, leading to rapid advancements in various related fields like big data, machine learning, and natural language processing. This has accelerated the overall pace of digital technology innovation, benefiting various industries.

2. A Change in Attitude Towards Digital Transformation: The excitement surrounding AI has served as a wake-up call for many business leaders, highlighting the urgency of digital transformation. This shift in attitude is crucial for businesses to remain competitive in the digital age.

3. Renewed Interest in Planning for a Fairer Digital Future: Concerns about the potential societal impact of AI have led to a renewed focus on planning for a fairer and more equitable digital future. This includes discussions about responsible AI development, investment in digital skills training, and the need for appropriate regulatory frameworks.

 When the Bubble Pops

For busy digital leaders and decision makers, navigating the AI hype and surviving the AI bubble requires a considered, balanced approach. Those I have seen make sustainable progress with AI adopt 3 main principles:

1. Focus on Business Value, not Hype: Don't get swept away by the hype. Instead, focus on identifying specific business problems where AI can deliver tangible value and ROI. Start with small, achievable projects and scale gradually based on results.

2. Build the Right Foundation: Address the fundamental challenges first. Invest in data infrastructure, talent development, and ethical considerations to ensure a smooth and successful AI implementation.

3. Collaboration is Key: Partner with experts, both internal and external, to bridge the knowledge gap and leverage the collective wisdom of your organization.

While the AI bubble might eventually burst, the underlying advancements in digital technology will continue to shape the future. Its significance lies beyond the hype. Leaders and decision-makers must focus on practical considerations, address challenges, and leverage the positive aspects of the AI phenomenon to steer their organizations towards a digitally transformed and ethically responsible future.

By focusing on real-world value, addressing challenges proactively, and embracing collaboration, digital leaders can harness the power of AI to drive meaningful change and achieve their digital ambitions. Remember, the journey is just as important as the destination. Navigate the hype with a critical eye, a pragmatic approach, and a focus on long-term value creation for your organization and society as a whole.