Human-perceived AI systems leverage the capabilities of technological intelligence (including existing AI models and AI languages) and human intelligence to achieve more than either could achieve on its own. Better performing AI models.
In other words, human-aware artificial intelligence is a kind of artificial intelligence enhanced by human intelligence. Some call it human perception technology, or human perception augmented intelligence. According to the latest research, artificial intelligence models of human perception appear to hold the key to incredible scientific breakthroughs, and over time, they may even reach AGI (Artificial General Intelligence) and ASI (Artificial Super Intelligence).
Jamshid Sourati and James Evans, the two scientists behind the study, agree that these AI models of human perception can accelerate scientific progress through the way they approach research methods. They trained unsupervised AI models on scientific research that only experts can recognize, while also making them aware of human expertise. The results were very promising: the AI improved predictions of future discoveries by 400%.
In other words, artificial intelligence models of human perception can make their own correct predictions about future scientific breakthroughs. These models succeed by predicting human predictions and the scientists who will make those predictions.
More importantly, artificial intelligence models of human perception are also able to distance themselves from existing theories and research methods, and they are able to come up with their own "alien" research methods that are now less common. May be pursued by humans. These methods generate scientifically promising hypotheses that can be further explored in research methods that have not yet been discovered or addressed by the human mind. AI models of human perception accelerate scientific progress by addressing blind spots and challenging current scientific limitations.
This may further accelerate the progress of artificial intelligence. We're talking about AGI and ASI, and big tech giants including Microsoft, OpenAI, Meta, and Amazon are currently paying attention. Augmenting human intelligence through artificial intelligence and vice versa, AGI can be easily achieved within this decade. Then reaching ASI is only a matter of time. According to the latest OpenAI project Superaligment, artificial superintelligence is within our reach. However, human-perception AI models will be the best type of ASI.
The above is the detailed content of Understand how artificial intelligence systems perceive human abilities. For more information, please follow other related articles on the PHP Chinese website!