AI’s productivity promise comes with mounting governance risks

Artificial intelligence is becoming a widely used tool for writing, research, education, programming and complex analysis, with nearly one billion people estimated to be using AI, according to the 2026 International AI Safety Report published by the United Kingdom’s government.

The gains are concentrated in particular tasks rather than the economy as a whole. The Bank for International Settlements reported in September 2025 that generative AI had delivered productivity improvements of 10 to 65 per cent and time savings of roughly 20 to 50 per cent in areas including coding, consulting and professional writing.

Businesses are also reallocating work around the technology. Indian technology company Wipro said AI had freed capacity equivalent to 20,000 employees, while it was training more than 100,000 workers in advanced AI skills. Demand is growing for machine-learning programming, chatbot development and software work that combines AI-generated code with human checking.

The employment effects remain uncertain but could be substantial. The International AI Safety Report estimated that up to 60 per cent of jobs in advanced economies and 40 per cent in emerging economies could be vulnerable, depending on adoption. It also identified suggestive evidence that early-career workers in occupations highly exposed to AI may be finding it harder to secure jobs.

The technology has increased risks alongside its usefulness. AI-generated text, voices, images and video can support deepfake scams and cyberattacks, while researchers warn that regulation and safeguards are struggling to keep pace. Anthropic chief executive Dario Amodei has called for slower improvements in AI capabilities to allow more time for careful use and stronger safeguards.

AI infrastructure is creating a separate policy challenge. Data centres require substantial electricity and water for computing and cooling, prompting concerns about energy supply, drought and local impacts. A Gallup poll cited by Al Jazeera found that 71 per cent of Americans opposed establishing AI data centres in their areas, while proposed projects in the United States and Kenya have drawn attention to their power requirements.