Will AI create or take away more job opportunities?

AI's impact on jobs is contested — I'll gather recent research to lay out both the creation and displacement side.

On balance, most large long-term forecasts say AI will create more jobs than it destroys - but in the short-term it will take away specific jobs, and the gains will be very uneven.

What the major forecasts say: Net gain

The most quoted forecast is the World Economic Forum's Future of Jobs Report 2025:

> while 92 million jobs might be eliminated by 2030, 170 million new roles will be created because of AI, resulting in a net gain of 78 million.

The WEF describes this as 14% of today's employment created vs 8% displaced.

An earlier version of the same WEF model for 2025 estimated:
> By 2025, AI might eliminate 85 million jobs but create 97 million new ones, resulting in a net gain of 12 million jobs.

Other projections point in the same direction:

The logic behind net creation is not just numbers, but type of change. The International Labour Organization's 2025 exposure index finds one in four workers worldwide is in an occupation with some exposure to generative AI, but only 3.3% of global employment falls into the highest exposure category. Its conclusion:

> because most occupations include tasks that still require human judgment, social interaction, responsibility, or context, "transformation of jobs is the most likely impact of GenAI".

The U.S. Bureau of Labor Statistics reached a similar conclusion: across 831 occupations, the BLS found no relationship between an occupation's AI exposure and whether employment is projected to grow or shrink. As one BLS economist noted: "It's not necessarily a determining factor in whether an industry is adding or losing jobs".

New jobs are expected to cluster around:

What is happening right now: Small net drag, uneven pain

MIT Technology Review's 2026 reality check argues:

> Despite the warning by some of an imminent jobs apocalypse that will destroy much of if not most such work ... there's scant evidence that AI has yet had any large-scale impact on the US labor market.

One reason: US Census data showing that only one in five companies are using AI in any business function.

Where displacement is measurable, it is concentrated and not yet offset:

Goldman Sachs payroll analysis for the 12 months ending March 2026 found:

> "In practice, we find that AI augmentation has created jobs, but not enough to fully offset the job losses from AI substitution, resulting in a small net drag on the labour market".

The bank estimated AI-driven automation eliminated about 25,000 jobs each month over the past year, while only about 9,000 were added back through productivity gains and new roles - a net loss of about 16,000 per month, or 0.1 percentage point of extra unemployment.

The most affected group is young entry-level workers:

In other words, coding and routine cognitive/language tasks are exposed, but experienced workers with tacit knowledge are still seeing wage gains - suggesting the "earn-while-you-learn" career ladder is breaking rather than the entire occupation disappearing.

Why both answers can be true

As economists from Stanford, Harvard and the Federal Reserve emphasize, there is no historical precedent for technology permanently reducing total employment, but there is also no guarantee this time isn't different - we are "flying blind" without better data on how AI is actually used in firms.

Bottom line: If you look at total number of jobs globally by 2030, most credible models (WEF, IMF, BLS, Goldman longer-term) forecast a net gain of tens of millions of jobs. If you look at who holds which job in the next 3-5 years, AI will take away more opportunities than it creates for people in highly automatable, entry-level knowledge work unless they reskill for AI-augmented roles.