By David Stephen

Consciousness science research of present-day is irrelevant, because on the two biggest stages in brain science: psychiatry and neurology, nothing from consciousness research is useful or applicable. There is no buzz around consciousness science because there is nothing promising. Even AI sentience research - which would be important to superintelligence, artificial general intelligence [AGI], AI safety, AI alignment, energy efficiency for data centers, solving hallucinations and so forth - is trashed. To make progress in consciousness science, with all its likely usefulness, a team would have to completely ignore everything in the field today and start afresh. Anything borrowed from existing theories or philosophy or whatever would doom it. OpenAI could be that team, given their success in exploding attention heads. To begin research in consciousness or for AI sentience, it is possible to start from another successful application today: GLP-1.

How will OpenAI impact on Consciousness and AI Sentience?

There is a recent spotlight from the School of Public Health at Brown University, A turning point in addiction psychiatry?, stating that, "These are GLP-1 receptor agonists. They're not just peptides like exenatide. They bind to specific receptors in the brain, especially in areas tied to the reward system: the ventral tegmental area, nucleus accumbens, prefrontal cortex. These regions regulate dopamine and motivation. By targeting these receptors, the drugs blunt dopamine release and reduce reward signaling. That means people feel less driven to seek out food, alcohol or drugs. That's the crux of how these medications may help with cravings."

What does it mean to not crave food? The question can be transplanted to, if there is a lesion, what does it mean to not feel pain from it? It could be possible to need food but not crave it. It is possible to be having a meal but not want more. It is possible there is a situation for pain, but to not feel that pain. How does the brain make these determinations?

A simple description could be that there can be a function, but the measure of that function may not reach a certain threshold to bear attention. So, while some situations may require attention, if the measure for attention is not attained, it may not get it. How does this explain consciousness? Functions have attributes that measure them. It is the collection of these attributes that can be defined, conceptually, as consciousness.

This explanation can be used to develop a test, or a standard to measure consciousness, in AI, non-human organisms and in humans. Questions of how functions are mechanized or how attributes come together can be theoretically explored. However, to narrow the objective for a standard, functions and attributes are targets.

Language

GLP-1s are as neuroscientific as anything can get. They are not philosophical or some disconnected theory of consciousness. If memory is a function and language is a subdivision of that function, what are the attributes that make language use possible? What attributes make language use seek outcomes of other functions - like emotions and feelings? Is language processed [much] differently from pain, pleasure, cravings or other subdivisions of feelings and emotions?

For humans, language use can be conscious. Language - speaking, listening, reading, writing, signing, singing - can be in attention or less than attention. Language can be subjective [or self or individualized]. Language also carries intent [or control or some free will], in how it is done properly. These attributes, attention, less than attention [or say awareness], subjectivity [sense of existence, as a chatbot] and intent are all attributes that ChatGPT and other AI chatbots use. For language alone, as a subdivision of a function [memory], it is possible to develop a measure for AI sentience. All the noise about AI is not sentient or AI consciousness is not possible would have to present a counter ...