The pursuit of artificial general intelligence (AGI), where an AI can perform tasks similar to a human, is at the forefront of research. This involves complex systems mimicking behaviors observed in natural organisms. Despite this, the belief that AI cannot obtain natural intelligence is prevalent. Some limitations of AI include its inability to navigate unpredictable environments and make decisions based on limited experiences. Therefore, the study of emotional responses in humans may enhance the AI’s ability, particularly examining the fear response.
Fear, governed by the amygdala within the brain, is considered a tool for keeping organisms safe from danger. The amygdala allows for a rapid defensive response once a threat is detected. Reactions to this threat include an increase in heart rate and blood pressure prepared for “fight or flight,” heightened attention and reflexes, and cognitive predisposition to a safe response. The amygdala, while primarily driving fear, is interconnected with other brain regions involved in perception, memory, and action.
Aside from humans, fear responses in insects, such as the common fruit fly, offer insights into primitive emotion. A study at Caltech demonstrated a suite of behaviors parallel to mammalian fear responses in fruit flies when exposed to an overhead shadow, indicating adaptive, self-preserving behavior. Understanding such responses within simpler organisms can therefore inform AI innovations.
The incorporation of fear into AI could enhance autonomous systems’ adaptability, such as self-driving cars. This relevance to safety arises from the need for rapid risk detection and responses. These fear-based responses are adaptable and can move beyond situations where the AI has been trained. Researchers are developing AI systems to emulate human defensive driving responses, via a fear-neuro-inspired reinforcement learning (FNI-RL) framework. Tests for the FNI-RL application proved successful, enhancing safety performance without reducing comfort or performance levels.
Bio-inspired AI, which applies neurologically and emotionally inspired responses to emulate humans, has expanded to sectors like manufacturing, healthcare, and space exploration. Memristors and Spiking Neural Networks provide bio-inspired chips to optimize AI systems for real-time learning and lower energy consumption. For example, the application of “digital nociceptors” in robotic hands would allow fast responses to damage.
In sum, research advocating for AI systems to utilize biological responses, such as fear, moves AI toward a new paradigm. Emotionally aware robotics may transform the way AI interacts with unpredictable environmental cues. This incorporation could lead to autonomous machines moving amongst us, demonstrating emotional reactions similar to humans.
