Unlocking Productivity: Why AI Needs a Human Touch to Solve the Puzzle

The conversation around AI and job losses often sparks two camps: those who worry about its impact and those who believe it could boost productivity. Microsoft’s CEO, Satya Nadella, suggests that AI can be a powerful tool, helping people set goals while it manages tasks efficiently. But history shows us that efficiency might not be enough to drive real productivity growth.

Over the past few decades, advanced economies have seen productivity growth fall from about 2% a year in the 1990s to around 0.8% recently. Even China, once a beacon of rapid productivity gains, has seen its growth stall. At first glance, advances in technology promised to revolutionize workplaces and spark a golden age of creativity. Yet, something seems to be missing.

Despite a wealth of information and global connectivity, scientific productivity has declined. Studies reveal that scientists are producing fewer groundbreaking ideas today compared to the 1960s. Why is that?

One theory is the quality-versus-quantity dilemma. Just like parents can’t give equal attention to every child, researchers may struggle to innovate when juggling multiple projects. Studies confirm that when researchers spread themselves too thin, breakthrough innovations become rare. Innovators like Isaac Newton understood this principle. He focused on one problem at a time, gradually finding solutions. Similarly, Steve Jobs emphasized that great innovation often comes from saying "no" to many ideas.

Real breakthroughs often happen in unexplored areas. If the 19th century had only improved existing tools like looms and plows, we wouldn’t have seen inventions such as antibiotics or rockets. Progress comes from discovery, not just speeding up existing tasks.

Large language models, like the ones used in AI today, tend to mimic existing knowledge rather than create new insights. They may help with routine tasks, but critical breakthroughs still rely on human creativity. Demis Hassabis, co-founder of DeepMind, notes that achieving true AI intelligence will require further innovation.

Currently, AI is great for efficiency. A recent survey of over 7,000 knowledge workers revealed that those who use generative AI can save about 3.6 hours a week on email tasks. However, this may not lead to lasting productivity increases. History, like the productivity boosts seen in the 1990s, shows us that without real innovation, new tools can only offer temporary gains.

For AI to genuinely enhance productivity, we need to focus on fostering creativity and originality. It’s about digging deeper into new possibilities instead of just refining existing ones. We should encourage bold ideas and give individuals the freedom to innovate. The technology is advancing; now we need to adapt our approaches to make the most of it.

For further exploration on the impact of technology on productivity, you might find insights in this OECD report on productivity growth.

Source link