Expert says human brains learn new information better than AI robots

We can find comfort in that human brains can still outsmart robots.

Though experts are concerned with the replacement of AI, as hundreds signed a letter warning of humanity’s 'risk of extinction' posed by AI.

As much as we're all worried about the powerful capabilities of AI, humans still hold the advantage in one key cognitive area.

Okay, so we may not match the speed of processing or provide instantaneous responses like AI, we humans excel in learning new information better than AI robots can, according to one expert.

Humans share an additional step that sets us apart.

Humans learn new information better than AI models / Jonathan Kitchen/Getty Images

A team of scientists at Oxford University analysed, simulated and compared 'information-processing models' of both human and AI models.

The team found that we humans employ a 'fundamentally different learning principle from that used by artificial neural networks.'

The study’s author Dr Yuhang Song at Oxford University said:

'Learning in the brain is superior in many critical aspects.'

When discussing AI capabilities, Song said: 'When trained on new tasks, performance on previous tasks is largely destroyed.

'Our biological systems often need to rapidly adapt to changing environments and have an advantage in this respect,' Song continued.

This important step reduces errors when we recall information, giving us an edge over AI systems.

The team discovered that a key advantage for learning in humans is their capability to 'consolidate new knowledge, skills, or memories.' This leads to fine-turning of nerve connections that result in reduced errors during memory recall.

AI models learn through 'backpropagation' /BlackJack3D/Getty Images

Unlike AI, which requires more exposure to learn and struggles with the interference of new and old information, the human brain's learning process is basically more efficient and robust.

The study, published in published in Nature Neuroscience, mentioned: 'We [humans] can learn new information by just seeing it once, while artificial systems need to be trained hundreds of times with the same pieces of information to learn them.'

Commenting on the study's results, Song said: 'For example, compared to the brain, backpropagation requires more exposures to learn and suffers from catastrophic interference of new and old information.'

AI learns through a process called 'backpropagation,' which is similar to the familiar process of elimination. Artificial systems will block errors until it finds an answer.

Song concluded from the study that existing computer simulations are relatively slow and called for the development of a 'new type of computer or dedicated brain-inspired hardware' that can rapidly implement 'prospective configurations' with little energy use.