Donnerstag, 23. März 2023

GPTs could upend labor market, new study suggests

My take on this is that sedentary head work with high education will be decimated. If robots are not flexible enough or too expensive then humans will keep their advantage as labourers of all sorts. The current political and social divide could be turned upside down. Good paying jobs will be sweaty labouring ones and those highbrow respected fields will be paid pennies on the dollar and left to ridicule. Primitive machos will regain social status. Cultural feminization, female empowerment will decline. Urbanization, office towers will reverse. If office work is replaced by a server then the cityscape could be flattened as with nuclear war effectively. Humans need activity and income for purpose. The permanent dream of a few hours work a week and Universal Basic Income is unlikely. I welcome a scifi novel take on this, perhaps written in 10 seconds by chat gpt explaining a possible future. 

https://cybernews.com/tech/gpt-upend-labor-market/

 The higher the entry requirements for a job, the easier it will be for AI to do it, a new paper from OpenAI indicates.

A new paper from OpenAI, published in collaboration with OpenResearch and the University of Pennsylvania, says that GPTs can “significantly affect a diverse range of occupations within the US economy.”


The paper is entitled “GPTs are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models” and was written, coded, and formatted with the help of GPT-4 and ChatGPT.


It argues that GPTs, or generative pre-trained transformer models, exhibit characteristics common to general-purpose technologies, also and somewhat confusingly sometimes abbreviated as GPTs.


General-purpose technologies are those that have a profound and far-reaching impact on society. Some innovations that upended the economic and social order in the past include the steam engine, electricity, and information technology.


Having assessed different occupations based on their correspondence with GPT capabilities, the paper concludes that these generative AI models could also have “notable economic, social, and policy implications.”


According to the paper, around 19% of workers in the US could have at least half of their work tasks exposed to GPT models, with the technology is expected to affect most occupations to varying degrees.


Interestingly, this impact will be felt more when it comes to higher-income jobs or those traditionally held by individuals with higher education and professional degrees. While professions such as blockchain engineers or interface designers could be “fully exposed,” occupations without any exposed tasks include short-order cooks, stonemasons, and mechanics.


“In other words, workers facing higher barriers to entry in their jobs tend to experience more exposure,” the paper says.


While “it does not necessarily suggest that their tasks can be fully automated by these technologies,” they will help workers save significant time in completing them.


Eventually, the extent to which AI models will be adopted will vary depending on the level of confidence humans can place in them, the study says.


“For instance, in the legal profession, the models’ usefulness hinges upon whether legal professionals can trust their output without resorting to verifying original documents or conducting independent research,” it says, noting existing risks associated with the technology, bias, and making up 'facts.'


It says the cost of adoption, the technology's flexibility, government regulation, and personal preferences will also play a role.


The paper was published by arXiv, an open-access research-sharing platform, and is not subject to peer review.

Keine Kommentare:

Kommentar veröffentlichen