Will AI replace over 50 million jobs by end of 2024?
➕
Plus
133
Ṁ47k
Jan 1
4%
chance

If an authoritative international labor or employment agency (like the International Labour Organization or the U.S. Bureau of Labor Statistics) releases a report stating that generative AI has replaced more than 50 million jobs globally by 11:59 PM Coordinated Universal Time (UTC) on December 31, 2024, the prediction will resolve as 'yes'. Any other outcome means the prediction is resolved as 'no'. Resolution is based on reports from these organizations.

Get
Ṁ1,000
and
S1.00
Sort by:

AI has the ability to both create new employment and opportunities while also automating some chores and responsibilities. Different industries and job kinds will probably see different effects of AI on employment. Some jobs might be automated, and AI might enhance some other jobs to increase productivity.

However, although many professionals and organisations are actively investigating and tracking the effects of AI on employment, precise forecasts are difficult to make with precision. The impact of AI on the labour market will also be influenced by societal variables such as worker training, policy choices, and others. All in all I think there are many jobs AI can replace, but many others are jobs only human can make, irresponsible.

@ErjolaAvdiaj I also bet NO due to the market question but that doesn't mean the 50 million replacements will not happend. According to the World Economic Forum, AI will replace 85 million jobs by 2025, but it will also create 97 million jobs (International Monetary Fund). Some of the most affected jobs will be accountants, customer service and secretaries. Also, according to a report by MIT and Boston University, AI will replace 2 million manufacturing workers by 2025 (Forbes). Furthermore, Goldman Sachs states that AI could replace 300 millon jobs and could rise global GDP by 7% over a 10 year period.

@ErjolaAvdiaj To deepen your analysis a bit more, let us review some of the predictions in this subject. According to an article of tech.co news, it is predicted that Generative AI will replace 2.4 million US jobs by 2030. The current question is providing a timeframe of one (1) year to generate 50 million jobs, with little to no human intervention. The truth is that given the eruption of AI, it is more likely that jobs will adapt and incorporate AI into their requirements, automating repetitive tasks.

In the same report, it is stated that Gen AI will influence 4.5 more jobs than it replaces, which refocuses the initial scenario of “replacement”. In figure 3 of that same report, it was forecasted that a total of 90,000 jobs would be lost by AI in 2023, where only 3,900 US jobs were indeed lost (seo.ai article). The point is that I think the question is assuming that the introduction of AI to the market will cause a lot of people to lose their jobs, when the tendencies and reports show that it will coexist and complement the current panorama.

Sources:

Report: https://regmedia.co.uk/2023/09/06/forrester_gen_ai_report_pdf.pdf

Article 1: https://tech.co/news/ai-replace-miilions-roles

Article 2: https://seo.ai/blog/ai-replacing-jobs-statistics#:~:text=In%20May%202023%2C%203%2C900%20US,were%20linked%20directly%20to%20AI.&text=British%20Telecom%20aims%20to%20replace%2010%2C000%20staff%20with%20AI%20within%207%20years.&text=Early%20AI%20and%20automation%20have,up%20to%2070%25%20since%201980.&text=Now%20it's%20Educated%2C%20white%2Dcollar,will%20be%20impacted%20by%20automation.

For reference: this is like 1-3% of employed people loosing or switching job, if I didn't mess up the calculation

predicts NO

When I bet this won't happen, I'm not making any claim about the relative pace of automation,

I'm saying the US Bureau of Labour Statistics et al won't frame the problem through this lens to measure the collected statistics in this way.

predicts NO

@AgenticLondoner This could be estimated by looking at all repots like "such and such news org fired x% of their editors because of AI" instead of by looking at employment rate stats.

But I'm not sure these agencies do such type of work.

As for looking at stats, if it's like 3% change, I'm not sure it's even above noise and needs to be explained.