Photo by The Hype and Reality of Artificial Intelligence: A Nuanced Perspective
As a long-time observer of the artificial intelligence (AI) landscape, I’ve witnessed the hype surrounding this technology ebb and flow over the years. From the promise of AI revolutionizing industries to the fear of machines surpassing human intelligence, the conversation around AI has been nothing short of polarizing. In this article, I’ll delve into the complexities of AI, separating fact from fiction, and provide a nuanced perspective on the current state of this rapidly evolving field.
Traditional AI vs. Generative AI: Understanding the Difference
To grasp the nuances of AI, it’s essential to differentiate between traditional AI and generative AI (GenAI). Traditional AI, which has been around for over 60 years, encompasses technologies like machine learning (ML), deep learning, and natural language processing (NLP). These technologies have been widely adopted in various industries, from healthcare to finance, and have proven their value in enhancing process efficiency and productivity.
GenAI, on the other hand, is a relatively new subset of AI that has gained significant attention in recent years. Born out of Google’s transformer technology in 2017, GenAI has made tremendous strides in generating human-like text, images, and even videos. The likes of OpenAI’s ChatGPT and other large language models (LLMs) have pushed the boundaries of what is possible with AI.
The Hype Cycle: When Enthusiasm Meets Reality
The AI hype cycle is a familiar phenomenon. Enthusiasm surrounding a new technology peaks, only to be followed by disappointment when reality sets in. Daron Acemoglu, institute professor at the Massachusetts Institute of Technology (MIT), aptly points out that while GenAI is a true human invention, too much optimism and hype may lead to premature adoption of technologies that are not yet ready for prime time.
A recent report by Gartner predicts that companies will shutter at least 30% of GenAI projects after proof of concept by the end of 2025 due to poor data quality, inadequate risk controls, escalating costs, or unclear business value. Gary Marcus, professor emeritus of psychology and neural science at New York University, goes a step further, predicting that the GenAI bubble will burst within the next 12 months, leading to an AI winter of sorts.
Measuring ROI: The Elusive Goal
One of the significant challenges in evaluating AI investments is measuring return on investment (ROI). As Andrew Ng, founder of DeepLearning.AI, points out, AI is the new electricity, transforming every industry and creating huge economic value. However, quantifying the ROI of AI is a daunting task, especially when it’s part of a broader delivery process.
Jayanth N Kolla, founder and partner of deeptech consultancy firm Convergence Catalyst, believes that seeking ROI is a disturbing trend emerging in the enterprise AI space. According to him, companies are making two critical mistakes: moving AI development from innovation and R&D budgets to software upgrade budgets and assigning stringent metrics and KPIs for AI solutions development and deployment.
Data: The Finite Resource
The success of AI models is contingent upon access to quality data. However, the finite nature of data sources like Common Crawl, Wikipedia, and even YouTube poses a significant challenge. Research from Epoch AI predicts that we will exhaust low-quality language data by 2030 to 2050, high-quality language data before 2026, and vision data by 2030 to 2060.
Don’t Put the Cart Before the Horse
In the rush to adopt AI and GenAI, companies often put the cart before the horse. Instead of identifying a business problem and finding the right technology to solve it, they are drawn to the shiny new toy of AI. As Andrew Ng points out, GenAI is absolutely safe enough for many applications, but not for all. It’s essential to take a step back, assess the problem, and then decide if AI or GenAI is the right solution.
Conclusion
The world of AI is complex, and the hype surrounding it is understandable. However, it’s essential to separate fact from fiction and take a nuanced approach to understanding the current state of AI. By recognizing the differences between traditional AI and GenAI, acknowledging the hype cycle, and focusing on the real challenges of measuring ROI and accessing quality data, we can make more informed decisions about the role of AI in our lives.
Illustration: The complexity of AI
In conclusion, while AI holds tremendous promise, it’s crucial to approach this technology with a clear-eyed understanding of its limitations and potential. By doing so, we can harness the power of AI to drive meaningful innovation and positive change in our world.
