Singapore's President Says AI Is Already Shrinking Entry-Level Jobs for Graduates

Two speeches at the IMF and World Bank annual meetings and one measurement from the New York Fed point to the same thing. The first job after graduation is in lower demand and often sits below the graduate's level of qualification.

On Monday 12 October 2026 in Bangkok, Singapore's President Tharman Shanmugaratnam said artificial intelligence is already reducing demand for entry-level jobs for graduates. In the United States, 42% of recent graduates aged 22 to 27 work in roles that typically do not require a degree^[1], according to the New York Fed.

Who spoke in Bangkok and who did the measuring

Tharman's remarks come from Reuters and were picked up by HCA Magazine. "There is a problem that has to be tackled very soon," the president said. "The first rung of the job ladder is seeing less demand today."

Tharman added that the evidence is clearest in the sectors most exposed to artificial intelligence.

On the same day, IMF Managing Director Kristalina Georgieva spoke of a "slow-moving tsunami" for jobs^[2], Bloomberg reports. CNBC^[3] also covered the meetings on the theme of artificial intelligence and work.

The measurement comes from the Federal Reserve Bank of New York and covers US graduates aged 22 to 27. The source reports it without the sample size or the collection method.

Unemployment holds steady while the work moves down a level

In the second quarter of 2026, unemployment among these recent graduates held at around 5.6%^[1]. Over the same period, the share employed in jobs that typically do not require a degree edged up to 42%.

These people are finding work, and they are finding it one rung lower.

The figure says where recent US graduates are working. It leaves the cause open. It applies to the United States and to that age group.

What the 40% estimate contains

Georgieva said the IMF expects artificial intelligence to affect around 40% of jobs worldwide over time. The estimate combines jobs that are enhanced, jobs that are downgraded and jobs that are eliminated.

It rests on IMF staff research published in January 2024, which breaks exposure down by type of economy.

• Advanced economies: around 60% of jobs exposed.
• Emerging markets: 40%.
• Low-income countries: 26%.

The same research estimates that in rich economies around half of the exposed jobs could benefit from the technology.

More recent IMF work looks at job postings. In a January 2026 article based on millions of online postings, Georgieva writes that one in ten postings in advanced economies asks for at least one new skill. Postings with a new skill pay around 3% more.

After five years, employment in the most exposed occupations was 3.6% lower in regions with strong demand for AI skills than in regions with weak demand. This is a correlation across regions.

The same analysis cites early US evidence that generative AI reduces entry-level hiring, particularly where tasks can be automated.

An estimate about the future and a measurement of today

The 40% is an estimate that the IMF places over time and that also includes jobs that improve. The 42% is a measurement already taken of people working today.

The gap between the two figures is the gap between waiting for a wave and reading your own hiring data.

Georgieva said the technology is already hollowing out demand for routine work. She added that lower-paid jobs are growing in sectors such as construction. In her view there is little to celebrate in that trade-off.

This newsroom's reading is that the entry-level role lost tasks before it was given new ones. Redesigning that role falls to those who lead the organization.

Who is opening the door to young people and what can be counted

HCA Magazine notes that the bank UOB intends to hire 1,000 young graduates through people-centered initiatives. The source gives the headline of the story and leaves out the results.

Three steps follow from the Bangkok numbers.

1. Count the entry-level roles opened this year and compare them with those of three years ago.
2. List the tasks that artificial intelligence has taken out of the first year of work and write down the ones replacing them.
3. Bring to the talent and compensation committee the share of newly hired graduates in roles below their qualification.

The 3% premium on postings with new skills gives whoever signs off the budget a first market benchmark for training new hires.

The question for the next board meeting

The question fits on one line. What work does a person in their first year at this organization do today, and who designed it?