The Future of AI to 2030: Forecasts vs Hype, Jobs and AGI

By the WahLiao desk · Last verified 3 October 2026

Nobody knows what Super Intelligence, the kind of AI behind ChatGPT, Claude and Gemini, will look like in 2030, us included. What we can do is sort the forecasts. The fairly safe bets: more capable agents, Super Intelligence built into most apps and phones (IDC expects about 70% of smartphones shipped in 2028 to be able to run it on the device), and cheaper access. The contested ones: human-level “AGI” by 2030, mass job loss, and a productivity boom. On jobs, employers surveyed by the World Economic Forum expect 170 million jobs created and 92 million displaced worldwide by 2030. On growth, serious estimates range from Goldman Sachs’ 7% lift to global GDP over ten years down to MIT economist Daron Acemoglu’s productivity gain of no more than 0.66% over the same period.

Our suggestion for a household: plan for the likely, not the loud. Build skills, keep a healthy dose of scepticism, pay for one tool at most, and treat any confident date for “AGI” as someone’s opinion, not a schedule. Every forecast below is named and dated so you can judge it for yourself, and if we’ve got something wrong, tell us and we’ll fix it.

AI to 2030: quick facts

Jobs (WEF, Jan 2025)	170 million created, 92 million displaced by 2030: a net 78 million, per employers surveyed
Jobs exposed (IMF, Jan 2024)	Almost 40% of global employment; about 60% in advanced economies; roughly half of those may benefit
Growth range	From no more than 0.66% productivity over ten years (Acemoglu) to +7% global GDP over ten years (Goldman Sachs)
Agents (Gartner, Jun 2025)	15% of day-to-day work decisions made autonomously by 2028, but over 40% of agent projects cancelled by end-2027
AGI by 2030	Google DeepMind’s chief puts it around 50/50; others, such as Yann LeCun, say today’s approach will never get there
Singapore	National AI Council chaired by the Prime Minister; four National AI Missions; over S$1 billion for public AI research and talent, 2025 to 2030

The WahLiao Verdict

Safe bet	On current forecasts, it’ll be in every app and phone you already own.
Coin toss	AGI by 2030. We wouldn’t plan your life around it.
Watch closely	Entry-level office jobs. The AI Index suggests the early squeeze is showing there first.
Take with salt	Anyone quoting one number for “jobs lost to AI”.
Do now	Learn one tool properly, pay for one plan at most, and double-check what it tells you.

What is reasonably expected

Agents, tools that carry out tasks rather than just answer questions (see AI agents explained), are improving quickly. The Stanford AI Index 2026 reports that on one computer-use benchmark, OSWorld, agent accuracy rose from roughly 12% to 66.3%, close to human level, yet agents still fail about one in three attempts on structured tests. Gartner forecasts that a third of enterprise software will include agentic features by 2028, while also predicting that over 40% of today’s agent projects will be scrapped by the end of 2027 for cost, unclear value or weak risk controls. Both can be true at once.

Super Intelligence in every app and device is the least controversial forecast we found. IDC expects 912 million phones able to run generative models on the device to ship in 2028, about 70% of the market. That doesn’t mean you need to rush out for a new phone; our guide to “AI” phones and laptops explains why.

Cheaper and more alike. The AI Index 2026 found the top models from several companies now sit within a narrow band on a popular head-to-head ranking, and says competition is shifting towards cost, reliability and usefulness. Our own read is that free and entry-level plans should keep getting better for the money, though that’s our guess, and no company has promised it.

Healthcare and education are where Singapore has committed to act. In schools, the Ministry of Education says Primary 1 to 3 pupils will not use chatbots for schoolwork, Primary 4 upwards will use Student Learning Space tools under supervision, and secondary students more independently.

What is contested

AGI timelines. Artificial general intelligence usually means a system as capable as a person across most tasks. It’s different from our house name, and from the still-hypothetical “superintelligence” far beyond human ability (see AGI and superintelligence explained). Google DeepMind’s chief executive, Demis Hassabis, has put the chance of AGI arriving within about five years, so around 2030, at about 50%, and at Davos in January 2026 said it may need “one or two more breakthroughs”. Yann LeCun, Meta’s former chief AI scientist, argues that today’s large language models will never reach human-level intelligence. When the field’s own leaders disagree this much, the honest answer is that nobody knows.

Mass job loss. The most alarming forecast came from inside the industry: in May 2025, Anthropic’s chief executive, Dario Amodei, told Axios that AI could wipe out half of all entry-level white-collar jobs and push unemployment to 10 to 20% within one to five years. The WEF’s employer survey points the other way, to net job growth, though 41% of employers said they plan to reduce staff where AI can automate work, and about two-fifths of workers’ core skills are expected to change by 2030. The early evidence sits somewhere in between: the AI Index 2026 reports that employment of US software developers aged 22 to 25 has fallen by nearly 20% since 2024, a squeeze on juniors rather than a collapse. For the Singapore picture, have a look at Budget 2026 and your job.

A productivity boom. McKinsey estimated in 2023 that generative models could add US$2.6 trillion to US$4.4 trillion a year in value and lift productivity growth by 0.1 to 0.6 percentage points a year to 2040. Goldman Sachs forecast a 7% rise in global GDP over ten years. Acemoglu’s estimate is far smaller. A spread that wide is a reminder that these are models built on assumptions, not measurements.

Forecasts to 2030, and how far to trust them

Prediction	Who says it	How confident	WahLiao’s read
70% of smartphones shipped can run generative models on-device by 2028	IDC, July 2024	High: hardware roadmaps already set	Likely. You’ll probably get it with your next phone anyway.
15% of day-to-day work decisions made by agents by 2028	Gartner, June 2025	Medium: analyst forecast	Plausible for routine tasks; keep the 40% cancellation forecast in mind too.
170m jobs created, 92m displaced by 2030	WEF Future of Jobs 2025	Medium: employers’ own expectations	A reasonable base case. Churn, not collapse.
Almost 40% of jobs exposed; 60% in advanced economies	IMF, January 2024	Medium: “exposed” is not “lost”	Singapore is an advanced economy, so expect your job to change.
Half of entry-level white-collar jobs gone, 10 to 20% unemployment in 1 to 5 years	Dario Amodei, Anthropic, May 2025	Low: an individual warning	Worth hearing, though we haven’t seen it in the data yet.
+7% global GDP over ten years	Goldman Sachs, April 2023	Low to medium: model estimate	The optimistic end of the range.
No more than 0.66% productivity gain over ten years	Daron Acemoglu, MIT, 2024	Low to medium: model estimate	The sceptical end. The truth may sit in between.
About 50% chance of AGI by 2030	Demis Hassabis, Google DeepMind	Low: depends on breakthroughs	A coin toss, by his own account.

What Singapore has committed to

These are commitments, not forecasts. The National AI Strategy 2.0, launched in December 2023, set a target to more than triple the number of AI practitioners here to 15,000. In the Budget 2026 statement on 12 February 2026, the government announced a National AI Council chaired by the Prime Minister, four National AI Missions in advanced manufacturing, connectivity and logistics, finance and healthcare, and a larger AI park at one-north. It also announced six months’ free access to a premium tool for people who complete selected courses (see the SkillsFuture offer) and a 400% Enterprise Innovation Scheme deduction on qualifying AI spending, capped at S$50,000 a year for Years of Assessment 2027 and 2028 (see grants for SMEs). On 20 May 2026, the Ministry of Digital Development and Information refreshed the strategy with ten priorities and noted over S$1 billion committed to public AI research and talent from 2025 to 2030. Schemes like these can be tweaked, so check the official page before you act. More in Singapore’s national plan.

How a household can prepare

Skills. Learn one general tool properly rather than dabbling in five. If you’re a citizen aged 18 or over, an eligible course now comes with six months of a premium plan. Our 90-day learning path is free.

Scepticism. The AI Index 2026 notes that a model won gold at the International Mathematical Olympiad in 2025, yet the best model on a clock-reading test read analogue clocks correctly only about half the time, against about 90% for people. So check anything that matters, and assume voice and video can be faked (see deepfake scams).

Savings on subscriptions. Free tiers cover most household needs. If you do pay, pay for one: Google AI Plus is S$6.98 a month, while Google AI Pro and Microsoft 365 Premium are S$28.99 (see plans in S$). Prices and plans change, so it’s worth a quick review once a year, perhaps when you’re going through your other annual bills.

Children’s education. Follow the school’s pace. Reading, writing and working problems out by hand still matter: they help a child judge when a machine is wrong. Our guide to children and Super Intelligence covers settings and ground rules.

AI in 2030: FAQ

Will AI take my job by 2030?

The forecasts we checked suggest it’s more likely to change it. The IMF says almost 40% of jobs worldwide are exposed, and roughly half of those may benefit; the WEF’s employers expect net job growth. Entry-level office roles are where the early pressure shows.

Will we have AGI by 2030?

Nobody knows. Google DeepMind’s Demis Hassabis puts it at about 50/50; Yann LeCun says current methods will never get there. Treat any firm date as one person’s opinion.

What is Singapore doing about AI?

A National AI Council chaired by the Prime Minister, four National AI Missions (advanced manufacturing, connectivity and logistics, finance, healthcare), a 15,000 AI practitioner target, and over S$1 billion for public AI research and talent to 2030.

Should I pay for an AI subscription to keep up?

Not necessarily. Free tiers cover most needs. If you do pay, pick one plan, or use the six months free after an eligible SkillsFuture course first.