Please consider supporting us by disabling your content blocker.
loader
Generative AI

Less than two years ago, the launch of ChatGPT sparked a generative AI frenzy, with predictions of a fourth industrial revolution reshaping our world. However, in March 2023, Goldman Sachs forecasted that 300 million jobs would be lost or degraded due to AI, indicating a significant shift was underway.

Fast forward eighteen months, and generative AI is not transforming business as expected. Many initiatives, including McDonald’s attempt to automate drive-through ordering, have been canceled after facing challenges.

The AI hype cycle

Generative AI has followed the Gartner hype cycle, which describes how initial successes lead to inflated expectations that often fall short. Following the peak of expectations, we enter a trough of disillusionment before reaching a plateau of productivity.

A Gartner report published in June indicated that most generative AI technologies are still at the peak or climbing, with full productivity two to five years away. Despite compelling prototypes, a RAND study revealed that 80% of AI projects fail, more than double the rate for non-AI projects.

Shortcomings of current generative AI technology

The RAND report highlighted challenges such as high investment requirements and a lack of skilled talent. Interestingly, generative AI systems can excel at complex tasks yet struggle with simple ones, leading to false confidence in their capabilities.

For instance, while they can solve complex equations, they may falter at taking straightforward drive-through orders. A recent study showed that large language models like GPT-4 often underperform in critical situations.

Why isn’t the generative AI hype over yet?

Despite challenges, generative AI technology is rapidly advancing, driven by scale and size. Research indicates that the size of language models and the data used for training significantly impact performance.

Large language models exhibit emergent abilities, showcasing unexpected skills as they reach critical sizes. AI companies continue to invest in larger models, with estimates suggesting generative AI must generate $600 billion in annual revenue to justify current investments.

What comes next?

As the hype subsides, we see more realistic AI adoption strategies. Companies are using AI to support human efforts, primarily focusing on improving efficiency and product quality. There is also a rise in smaller, cost-effective generative AI models tailored for specific tasks.

Moreover, a strong emphasis on AI literacy training is emerging, helping the workforce understand AI’s capabilities and limitations. The AI revolution is likely to evolve gradually, enhancing human activities rather than replacing them.