It would be nice if innovation could happen "just by dreaming," says Viktor Mayer-Schönberger, who stresses that imagining things is "crucial for being innovative."

That is why the professor of information science at Oxford University believes artificial intelligence (AI) won't be able to compete with human creativity for the time being.

"Humans can imagine things that don't exist yet," he told DW, because despite being trained on massive datasets AI is working with data from the past.

So, the datasets used for machine learning reflect what we can learn from the past for the present, he added, enabling AI to make insights from collected data more accessible, "but it doesn't invent anything new."

If the present or the future differs significantly from the past, AI cannot help us find the right solutions. For example, if people during Henry Ford's time had been asked what they wanted, most would have likely said "a faster horse" — a solution rooted in past experiences.

AI is thus a tool for evaluating large datasets and increasing efficiency, especially in economically stable times. However, we do not live in stable times. The challenges posed by climate change require innovations beyond current capabilities.

Paradoxically, the pace of innovation has slowed down despite rapid advancements in AI, said Mayer-Schönberger.

Ufuk Akcigit, an economics professor at the University of Chicago, and Sina T. Ates from the board of the US Federal Reserve both have observed a slowdown in productivity growth in the United States.

"The entry rate of new businesses has decreased, productivity growth has slowed down, the labor share of output has decreased, while market concentration and the corporate profit share of Gross Domestic Product (GDP) have increased," they wrote in a recent paper on declining US business dynamism.

The researchers found that the dynamism of innovation in American companies has decreased since the 1980s, and even more noticeably since the 2000s.

They attribute this to insufficient competition between leading companies and their rivals, partly because knowledge is not shared sufficiently. This prevents latecomers to markets from learning from the advances of the leaders and growing themselves. As a result, there is less competitive pressure on the big players, who, without competition, have fewer incentives to innovate.

The prime mover of modern-day innovation is data. With the help of AI, large amounts of data can be increasingly well-analyzed.

According to the Federation of German Industries (BDI), ever more data is being collected, with the volume having increased 10 times between 2012 and 2022, and expected to triple again by 2025.

This is where major digital companies like Google, Amazon, and Facebook come into play. These companies collect vast amounts of data, thus becoming more efficient while at the same time preventing access to their data wealth for others.

"Although these digital giants are reputed to be pioneers, they actually slow down innovation processes and progress by hoarding data," said Mayer-Schönberger, adding that rival companies, as well as public institutions and scientific organizations, are shut out.

To view this video please enable JavaScript, and consider upgrading to a web browser that supports HTML5 video

These days, it is also more common for innovative companies to be simply bought out by big firms, said Mayer-Schönberger. He noted that about 20 years ago more than three-quarters of successful Silicon Valley startups chose to go public, while today, three-quarters are already swallowed up by the likes of Google and Facebook even before they go public.

The Oxford University researcher warns this is not only hampering innovation but is also presenting a systemic risk to economic growth. He likened the situation to driving on a highway where if one car's brakes fail, it's bad enough, but if all the cars have the same faulty brakes because there is only one type of car, it leads to a crisis.

This dilemma can only be resolved, he said, if policymakers ensure that data is more freely accessible.

"There is no property right to data," Mayer-Schönberger said, like "intellectual property rights, authorship, and patent rights produced through human thought." Data, however, "cannot establish ownership," he added.

While the EU's Digital Services Act and the Digital Markets Act are steps in the right direction, Mayer-Schönberger said, especially education systems must also be restructured to provide the next generation with "incentives to dream purposefully."

"It's not about memorizing a 19th Century poem but about seeing the world with different eyes," he said, because innovation doesn't need "streamlined, diligent ants, but uncomfortable mavericks."

This article was originally written in German.

QOSHE - Are Google, Meta and Amazon's AI tools hampering innovation? - Insa Wrede
menu_open
Columnists Actual . Favourites . Archive
We use cookies to provide some features and experiences in QOSHE

More information  .  Close
Aa Aa Aa
- A +

Are Google, Meta and Amazon's AI tools hampering innovation?

55 1
29.05.2024

It would be nice if innovation could happen "just by dreaming," says Viktor Mayer-Schönberger, who stresses that imagining things is "crucial for being innovative."

That is why the professor of information science at Oxford University believes artificial intelligence (AI) won't be able to compete with human creativity for the time being.

"Humans can imagine things that don't exist yet," he told DW, because despite being trained on massive datasets AI is working with data from the past.

So, the datasets used for machine learning reflect what we can learn from the past for the present, he added, enabling AI to make insights from collected data more accessible, "but it doesn't invent anything new."

If the present or the future differs significantly from the past, AI cannot help us find the right solutions. For example, if people during Henry Ford's time had been asked what they wanted, most would have likely said "a faster horse" — a solution rooted in past experiences.

AI is thus a tool for evaluating large datasets and increasing efficiency, especially in economically stable times. However, we do not live in stable times. The challenges posed by climate change require innovations beyond........

© Deutsche Welle


Get it on Google Play