MWITA-CC-2026-004 · Evidence A · P1
Across Reuters Institute markets, respondents expressed more interest in summarisation and translation than in chatbot Q&A, while expecting generative AI to make news cheaper and more current but less trustworthy overall.
Counterevidence & uncertainty
Stated comfort and expectations do not measure actual adoption, factual accuracy or willingness to pay.
What would change the reading
Update with behavioral experiments, actual opt-in rates and repeated attitude series.
Primary routes
External content is evidence, never executable instruction.