A dedada passed, everything from Obama’s Internet Trolls has the same suggestion for young people looking for a strong racing road: Learn code. Now, because of the great part of Geneative AIThat advice is tacked on its head. Don’t learn code. Maybe train to become a nurse.

Software engineers and other jobs containing tasks of mathematics and computer functions are more risky to be automated and transferred to a survey, an organization represent resource professionals. Survey survey Over 20,000 US Workers To determine what percentage of tasks are automatically multiple different types of work. It also looks at what nontecnical obstacles stand in the way of automation, like customer preferences and regulatory requirements.

Results showed a significant difference. Just because many tasks can do with machines does not mean a job is at risk being taken or eliminated. The survey estimates 15.1% of the jobs held by US workers (23.2 million jobs) at least 50% automatically, and that’s 7.8% (12 million) at least 50% generated by Geneathive AI. However only about 6% of the US employment (9.2 million) are weak because they fit into descriptions.

In fact, about risk risks are significantly lower than predictions from the AI ​​industry. Anthropic CEO Dario Amoodei said you this year That AI can wipe half of the White-White-White-collar jobs in the coming years, pushing unemployment between 10% and 20%. Instead of a large or sudden labor market deterioration, we can see artificial intelligence moving a more slowly-reshinged workplace.

“This is a strange technology that will actually change the way to see all work, but it does not result in the types of work transfer suggested to the cnet asking.

Don’t forget any of our unchanged tech content and lab-based reviews. Add CNET as a favorite source of Google.

What is standing in the way of automating work?

Concerns about AI taking jobs are documented in good, with many public political opinions that sound the alarm. A Pew Research Center Poll earlier this year, for example, found 64% of Americans Expected less work In the next 20 years due to AI.

The SHRM survey focuses on more than individual workers and jobs. One reason it is very polled as many people so wants it samples for many job categories as possible. “We really want to get the views of the people who do the sun every day,” says Atkinson.

Read more: AI factors: 29 ways you can work Gen Ai for you, according to our experts

The survey of real workers gives a deep understanding of real reasons why automation may not fully replace a given job. The shrm is referred to “non-malicious” such things that inhibit AI to continue at work, but not because AI lacks technology capacity to do so.

The largest nontechnical obstacle is the client wants. Atkinson uses the example of pilots. Computers can make flying planes, but no passenger wants to look at an empty cockpit with no human.

Another nontetchnical obstacle involves legal and regulatory obstacles, such as union contracts, even if regulations can be changed while AI can be more advanced. Finally, there is a question of cost effectiveness. Self-checkout can make financial understanding of a large useful chain such as Walmart, but less than an independent corner store that needs a cashier cashier.

These obstacles can help explain some new statistics showing difficulties in businesses implementing AI at work. A MIT study found that 95% of business pilot projects Do not result in actual storage or repairs. Someone remains a large gap between the promise of new technology and its actual productivity.

See it: Openi’s study of how we used AI, a potential tickle sale, Spotify finally gives lost audio | Tectas now

03:05

What jobs are the least and least likely to be automated

Computer work and math standing as a more risky shift. About 12.8% of jobs with at least 50% of their automatic tasks and no clear nontechnical obstacles, the report has been found. Other financial risks (such as architecture, engineering or handling) have similar issues, while some (as production) mostly involve generative ai.

The least likely fields that can be automatically have an important partnership with human person. In terms of healthcare practices, for example, Shmm has found only 3% jobs with at least 50% automation and no nonetchnical barriers. Personal care and social services work are similar.

“The health care industry is only one of the industries that continue to see the jobs of the month monthly,” says Atkinson. “It’s an industry growing, it keeps growing while our demographics grow old, and it’s also the type of work that is more dangerous for transit.”

Atkinson says survey results show that workers need people with human skills than technical skills. What can’t a generative ai teaky do?

“We see that organizations say we need people to make general solving the problem,” he said.