Despite recent advances in the field of artificial intelligence (AI), it is undeniable that current models still have significant limitations. However, a group of Chinese scholars from the Hong Kong Polytechnic University (PolyU) have taken a step towards overcoming these barriers by demonstrating that AI can be trained to think similarly to the human brain.
The research, conducted by the PolyU team, suggests that training AI models similarly to language processing in the human brain it could revolutionize not only the AI sector, but also our understanding of how the brain works itself. This innovative approach is based on the fact that each AI model is essentially linguistic and is trained using data of this type.
Currently, large language models (LLMs), such as ChatGPT, are mainly based on contextual word prediction for their functioning. However, researchers have explored an alternative approach, known as next sentence prediction (NSP)which aims to simulate the process of understanding speech at the brain level.
The results of the study were surprising: the model trained with the NSP showed a much closer correspondence with human brain patterns
, especially in areas of the brain involved in understanding speech. This model was able to better replicate human brain activity and more accurately predict people’s reading speed.
The implications of these findings are significant and suggest that NSP can make AI models more efficient and similar to human thinking. However, it is important to note that training these models requires the use of a huge amount of synthetic and non-synthetic data.
With these advances, we are approaching a moment in history where AI could gain astonishing, almost human-like reasoning abilities. This development will not only open up new frontiers in the field of artificial intelligence, but could also provide valuable insights into how the human brain works.
As research proceeds, it is critical to consider the ethical and social implications of AI increasingly resembling human thinking. In-depth debate and appropriate regulation will be needed to ensure that these powerful tools are used responsibly and for the benefit of humanity.
