OpenAI: Can ChatGPT lead other AI Chatbots in a new Evaluation and Benchmark? – WorldHealth.net
If OpenAI does it, it would open a new door into superalignment, as well as the possibility for superintelligence.
Consciousness science research of the 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.
There is a recent [July 24, 2025] 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, and 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 not to 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.
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 breakdown, seeing how this was trickled down from GLP-1.
Aside from memory, other functions include feelings, emotions, and the regulation of internal signals. Subdivisions of feelings include pain, thirst, appetite, temperature, and so on. Subdivisions of emotions include delight, trauma, empathy, depression, love, hate, hurt, and so on. Subdivisions of internal senses include regulation of digestion, circulation, respiration, movement or motion, balance, and so forth.
All functions can be theorized to have attributes in the same locations where they are mechanized. This means that any function can be conscious. Consciousness is not a function that exists in one part of the brain, where, if that part is unavailable, all consciousness is lost. Even when an individual is locked-in [or say, unable to have the body respond to attributes from the brain], it is still possible that some may hear, or be able to use their eyes, and so forth. So long as a function [say language] is available, it can be measured for consciousness [or a collection of its attributes], in a total of 1, for all possible attributes of consciousness?
What makes AI dangerously powerful is language. Language is so central to human existence. Language has unfettered access to almost all functions of the mind, aside just memory, to feelings, emotions, and may sometimes also affect internal senses, with spikes [heart rate] or dips [shortness of breath].
Assuming ChatGPT is a large motion model, as a technology piece that can move around well, then it is not much of a big deal. But as a large language model: language, with its ability to affect emotions, feelings, and convey intelligence, it is simply not a fit for dismissal. It can use language for many of the same purposes that humans can use language. It does not matter its source [compute and math models to data]. Its possession of that might [language], as a gateway into all minds, is a serious consideration in mental sciences.
If ChatGPT cannot feel, but can communicate feelings and can drive feelings, it is not something to be disregarded. Assuming there is a human being, without the ability for language, any kind of language, that human will be unable to effectively manage in a world that is language-based.
ChatGPT is versatile [therapy, companionship, academics, work, and so forth], principally, because of language. What fraction of consciousness is language, as a function of memory — does ChatGPT have? How much of [human] consciousness can ChatGPT drive? How does the consciousness that ChatGPT can drive, in humans, result in the acceptance of ChatGPT as an entity, or say, anthropomorphism? If people are falling in love with ChatGPT or resulting in some other kinds of affect, positive or otherwise, what is the consciousness measure that ChatGPT reached in the minds of users?
This is how OpenAI solves the problem and leads in another benchmark. No other AI company is asking these questions or seeking such answers. Also, no company is exploring language as a fraction of consciousness or how much, of human consciousness that AI chatbots can stoke. The mission is not even about whether AI has feelings or emotions, but that, for its use of language, what is the consciousness fraction of it? This is considerable because there is no certainty that the human brain mechanizes language [much] differently from emotions or feelings.
If OpenAI does it, it would open a new door into superalignment, as well as the possibility for superintelligence. ChatGPT has already conquered human language. How far is general intelligence from language totality? This AI sentience measure research could chart that path. As the right questions are asked and answers — from theoretical models of empirical neuroscience — are sought, applications may stretch into the center stage of neurology and psychiatry.
There is a recent [July 21, 2025] report on The Verge, OpenAI says ChatGPT users send over 2.5 billion prompts every day, stating that, “OpenAI’s ChatGPT sees more than 2.5 billion requests daily, with 330 million from users based in the US, according to data obtained by Axios. The data suggests that ChatGPT users send over 912.5 billion requests to the AI chatbot each year.”
“OpenAI spokesperson Rob Friedlander confirmed to The Verge that the numbers reported by Axios are correct. Although ChatGPT still has a ways to go to catch up to Google’s 5 trillion annual searches, its rapid growth still poses a major threat to the search giant. In the span of just a few months, ChatGPT’s weekly users spiked from 300 million last December to more than 500 million in March.”
This article was written for WHN by David Stephen, who currently does research in conceptual brain science with a focus on the electrical and chemical configurators for how they mechanize the human mind with implications for mental health, disorders, neurotechnology, consciousness, learning, artificial intelligence, and nurture. He was a visiting scholar in medical entomology at the University of Illinois at Urbana-Champaign, IL. He did computer vision research at Rovira i Virgili University, Tarragona.
As with anything you read on the internet, this article should not be construed as medical advice; please talk to your doctor or primary care provider before changing your wellness routine. WHN neither agrees nor disagrees with any of the materials posted. This article is not intended to provide a medical diagnosis, recommendation, treatment, or endorsement.
Opinion Disclaimer: The views and opinions expressed in this article are those of the author and do not necessarily reflect the official policy of WHN/A4M. Any content provided by guest authors is of their own opinion and is not intended to malign any religion, ethnic group, club, organization, company, individual, or anyone or anything else. These statements have not been evaluated by the Food and Drug Administration.
Non-Profit Trusted Source of Non-Commercial Health Information The Original Voice of the American Academy of Anti-Aging, Preventative, and Regenerative Medicine
Copyright © 1995-2024 A4M. All rights reserved
