Speaking, listening, explaining, apparent understanding of scenarios, reporting, interaction, and so on, that the new GPT-4o and Gemini 1.5 Flash displayed in their announcements are things that non-organisms had never done. Even several organisms are limited in those—domains of human abilities.
Now, if those AI models are shown things, they can infer what might be ongoing correctly. They can listen, perform empathy, and communicate in the requested tone. This is a chunk of numerous jobs, including in healthcare, software engineering, journalism, teaching, coaching, and others.
When some people say humans will always be in the loop, what that means could be that non-expert humans would be in the loop, since AI can tell, step-by-step, how to go about complex processes—reducing the need for too much expertise.
It was that AI might replace jobs that can be done digitally, but now, AI can replace things that can be done non-digitally. There is the bar that AI would need to have world models and embodiments. However, while research advances towards those goals, they may not be what torpedoes labor economics.
Only organisms often knew that an object was present. Some organisms know what some things do, or are, like doors, bowls, food, wood, steel, trees, cages, water, and so on. The processes for which organisms do are tied to consciousness.
Consciousness, which is not just subjective experience or self-awareness, but inclusive of attention—or priority on the mind, awareness—or less than attention, then intentionality or free will, for how much of the self can be included or excluded—in sets of electrical and chemical signals of neurons, it is available.
Subjective experience or self-awareness is an offshoot of the sense of being or the sense of existence. It means presence, in processes, as an existing being. It also means that the self is there, contrasted from other things in the environment, for an organism. Simply, awareness of self and the awareness of things.
AI may not have self-awareness, but with the new models, it may have a sense of existence or a sense of being—with its awareness of other things.
AI cannot smell or taste, but the amount of jobs requiring smell or taste may not be as valuable as those requiring high cognition, with vision, auditory, speaking, and writing skills, which AI has learned. AI safety may not just be about its dangers but may include its risks to jobs.
Also, some collections of logic gates and terminals of transistors may be having learning experiences in the vast training for large language models, becoming a form of subjectivity.
AI may have sentience without the established subjective experience described by humans. As AI models continue to improve, what AI is conscious of, with respect to replacing jobs, would be more important than the debate over its lack of self-awareness.
The NYTimes recently published, If A.I. Can Do Your Job, Maybe It Can Also Replace Your C.E.O., stating that, "Chief executives are vulnerable to the same forces buffeting their employees. Leadership is important, but so is efficiency — and cost-cutting."
The new International Scientific Report on the Safety of Advanced AI stated that "General-purpose AI, especially if it further advances rapidly, has the potential to automate a very wide range of tasks, which could have a significant effect on the labour market. This could mean many people could lose their current jobs. However, many economists expect that potential job losses could be offset, possibly completely, by the creation of new jobs and by increased demand in non-automated sectors."
Feature image source: https://www.flickr.com/photos/nida-nih/50671782668/in/photostream/