Nearly Three-Quarters of Adults in US Believe AI Will Eliminate More Jobs Than It Will Create -- and They Want Companies To Pay For the Retraining (gallup.com)

Key findings from a Gallup poll: Nearly three-quarters of adults (73%) say an increased use of AI will eliminate more jobs than it creates (PDF). Results are consistent across most demographic groups. However, those with blue-collar jobs are particularly pessimistic, with 82% saying the transition will result in a net job loss, compared with 71% of those with white-collar jobs.

Nearly half of Americans (49%) say "soft" skills, such as teamwork, communication, creativity and critical thinking, are the most important for U.S. workers to cultivate to avoid losing their jobs to AI. Alternatively, 51% say learning "hard" skills, including math, science, coding and the ability to work with data, are the most important to maintain a job in the face of new technology adoption.

When asked to choose among seven options concerning who should pay for retraining, a clear majority of U.S. adults overall (61%) say employers should fund these programs. The federal government comes in second at 50%.

Comments

Well the good news is 3/4 figured it out

[rsilvergun]

the bad news is 1/4 haven't. When the industrial rev took off it put more folks out of work than it employed. That's where Luddites came from. The were freaking out over losing their livelihoods in a society where your quality of life is determiined by your job.

It took 80 years for other tech to catch up and employee more people than it put out of work. The people alive during those 80 years either lived like kings or like crap. And as far as I can tell nothing's changed. Your quality of life is determined by your job (or lack thereof)

10 things AI won't do

[sycodon]

1. Change the clutch on my car.
2. Fix my home's AC.
3. Trim my trees.
4. Talk to me about my investments.
5. Diagnose my illness (without a doctor as the interface)
6. Teach my kids.
7. Police my neighborhood.
8. Put out a house fire.
9. Rescue someone.
10. Get elected and participate in government.

AI is a tool that could help with all this, but it isn't a thing that can do all of this.

Re:10 things AI won't do

[Monster_user]

1. Change the clutch on my car.
- An AI in a robotics shop will be able to assemble and disassemble vehicles. This is kind of where AI began, with Ford's assembly line. Eventually changing the clutch won't be a problem. Older models will be considered hobbyist, not job creating work.

2. Fix my home's AC.
- How many different ways can an AC break? The question isn't whether an AI powered robot can fix AC, but whether it is cost effective to do so.

3. Trim my trees.
- Is this a matter of art and presentation, or function? Safety would be a concern, we are a long way off from feasibly being able to turn an AI driven "killer robot" loose in the yard and expect it to only trim trees.

4. Talk to me about my investments.
- Simple investment advice based on trends and spreadsheets? AI's cheap and effective. The ability to teach you about your investments and/or investing is something AI may never be able to do. At any rate, it would require very advanced AI to be able to speak in layman's terms.

5. Diagnose my illness (without a doctor as the interface)
- AI is already making progress in this area. AI doctors are on their way. At the very least this reduces the need for highly skilled medical professionals, as those fewer professionals would be able to focus on outliers or challenging illnesses.

6. Teach my kids.
- Teach your kids what? This is one of the targets for AI, getting it to pass a Turing test. Children typically have simpler questions for which there is usually adequate knowledge and refinement to turn into an AI. Most of what your kids learn in school can be taught by AI.

7. Police my neighborhood.
- I'd agree with this. Policing is a very complex and intuition heavy job, very centered on the human experience. Basic surveillance, such as speeding ticket cameras, and home security systems are probably a significant portion of AI's abilities in this area.

8. Put out a house fire.
- AI ought to exceed here, outside of the first responder issue. Too many variables between a fire station and a fire. If you can get an AI and equipment to a house fire, then it can make better decisions on how to reduce or minimize damage to your house and belongings. Potentially an AI with the proper sensors would better be able to rescue trapped pets and humans without risking loss of life of first responders. However, due to the lack of savings an AI and robotics combination will bring, I doubt any use of AI in the firefighting arena to affect jobs anytime in the near future.

9. Rescue someone.
- Back to the above. Depends on the circumstances, and the risk to bystanders. Rescuing individuals often entails going "outside the script", or encountering circumstances which are "outside the script". However, there may be many types of risk which can have programmed mitigations, or nearly foolproof fallback plans. The more we can system-ize something, the greater the possibility we can turn that system into an AI program.

10. Get elected and participate in government.
- Actually, I see this as a possibility. There would be some concerns as to what areas AI would be safe to govern, not everything should be a simple pre-programmed response. And learning AI is about implementing what we would "pre-program" into the AI given infinite time and resources.

Re:tax them

[ShanghaiBill]

we need to start taxing companies who use AI/robots that take away jobs

Exactly. We need to start by taxing everyone that uses a washing machine or dishwasher rather than employing a laundress and scullery maid.

Re:They're not wrong

[Anonymous Coward]

;In the short term, we're in for epic disaster levels of unemployment. Only the owners of capital will be immune to the worst effects.

Of course, in the long term the economy will adjust and we'll use our extra productivity to sell each other goods and services we previously wouldn't have bothered with...

That's where you're wrong. In the long term, the results will be even worse.

A hundred years ago, when Henry Ford began producing automobiles, he paid his workers higher wages than most other companies of that time. He didn't do it because he was a nice guy, he did it because he recognized a very simple and basic reality: those people aren't just workers, they are customers. Higher wages means he sells more cars.

Contrary to the bullshit "trickle down economics" that Republicans have pushed for 35 years, the truth is, our economy is based on trickle-up economics. When you give people at the top more money, they keep it. But when people at the bottom have more money, they spend it and the money flows up to the business owners and that's how they become wealthy.

And that's the big problem that nobody is talking about. Yes, in the short term the rich get richer and everyone else gets poorer. But, ironically, in the long term, everyone gets destroyed, including the wealthy, because of that one simple reality: Once you've eliminated all the jobs and 98% of the population is living in poverty, who will buy your products?