MIT and University of Queensland Warn of Top AI Risks as Global Entry‑Level Jobs Slip 29%

The MIT FutureTech laboratory and the University of Queensland released a joint study that ranks the five AI risks most likely to cause catastrophic outcomes within the next five years. Using a Delphi survey of 272 experts, the report identifies dangerous capabilities, intense competitive pressure, AI‑enabled weapons and cyber‑attacks, power centralisation, and false information as the highest‑priority threats. It highlights the vulnerability of the information sector, national‑security agencies, and the finance‑insurance industry, noting that many of the 24 risks assessed have over a 10 % chance of leading to more than one million deaths or $100 billion in economic loss.

At the same time, new labour‑market data shows that global entry‑level white‑collar positions have fallen 29 % since early 2024 as AI automation spreads. Young workers aged 22‑25 have seen employment rates drop by 2.7 %, with larger declines in finance, software and creative fields. The OECD points to a pronounced impact on the UK service sector. AI systems now handle complex tasks ranging from code generation to financial analysis, prompting firms to adopt AI rapidly while also imposing usage limits to control rising computational costs. Experts stress that AI is reshaping rather than wholly replacing many jobs.