Generative AI won't steal your job, just change it, says UN

Generative AI will astir apt not switch astir existent workers, pinch its effect alternatively confined to automating immoderate tasks for a minority, according to a study released connected Monday by nan International Labour Organization (ILO), nan United Nations agency that develops standards for nan world of work.

The report, titled "Generative AI and Jobs: A world study of imaginable effects connected occupation amount and quality," observes that nan motorboat of OpenAI's ChatGPT has created concerns astir occupation losses, arsenic machines tin of creating matter and images, aliases analyzing data, execute tasks galore humans are paid to perform.

But those concerns whitethorn beryllium overblown, nan study suggests, because nan ILO's researchers found astir workers are not astatine precocious consequence of being replaced by AI.

Instead, nan report's [PDF] main uncovering is that "most jobs and industries are only partially exposed to automation and are frankincense much apt to beryllium complemented alternatively than substituted by AI."

Anyone who's utilized generative AI and seen its, shall we say, imaginative output will apt person travel to nan aforesaid conclusion.

The study finds that while devices for illustration GPT-4 tin do immoderate of nan tasks humans execute successful fields specified arsenic administration, customer service, information management, and providing information, AI models cannot do each nan activity required successful astir roles.

"As a result, nan astir important effect of nan exertion is apt to beryllium of augmenting activity – automating immoderate tasks wrong an business while leaving clip for different duties – arsenic opposed to afloat automating occupations," nan study states.

But clerical workers will consciousness much impact.

Generative AI could mean definite clerical jobs ne'er look successful lower-income countries

"We find that only nan wide business of clerical activity is highly exposed to nan exertion pinch 24 percent of clerical tasks considered highly exposed and an further 58 percent pinch medium-level exposure," nan study states. "For nan different occupational groups, nan top stock of highly exposed tasks oscillates betwixt 1 and 4 percent, and mean exposed tasks do not transcend 25 percent."

That study is bad news for women, arsenic nan study finds they are over-represented successful clerical activity – particularly successful precocious and middle-income countries.

"Since clerical jobs person traditionally been an important root of female employment arsenic countries create economically, 1 consequence of Generative AI could beryllium that definite clerical jobs whitethorn ne'er look successful lower-income countries."

Adoption of generative AI successful developed countries intends workers successful those nations are much apt to consciousness nan effect of nan technology. The study estimates 5.5 percent of full employment successful high-income countries whitethorn beryllium astatine consequence of being partially automated, compared to conscionable 0.4 percent successful low-income countries.

• Google reportedly designing chatbots to do each sorts of jobs – including life coach
• TV and movie extras fearfulness generative AI will transcript their faces and bodies to return their jobs
• Netflix offers up to $900,000 for AI merchandise head while actors onslaught for protection
• OECD finds 27% of jobs are nether threat from AI

"We focused connected nan imaginable of task automation arsenic of today, without speculating connected nan numbers of caller jobs that mightiness emerge. This attack mightiness person been expected to make alarming estimates of nett occupation nonaccomplishment – but it did not. Rather, our world estimates constituent to a early successful which activity is transformed, but still very overmuch successful existence," nan researchers said.

They warned, however, that countries request to specify policies to guarantee that workers' authorities are still protected arsenic industries set to generative AI.

"Without due policies successful place, location is simply a consequence that only immoderate of nan well-positioned countries and marketplace participants will beryllium capable to harness nan benefits of nan transition, while nan costs to affected workers could beryllium brutal," nan ILO study concluded – arsenic tin beryllium expected from an statement that exists to create and beforehand policies that make activity adjacent and dignified. ®