20+ AI Privacy Statistics (2025)

of people globally are willing to trust AI systems, less than half the world's population

KPMG's 48,000-person, 47-country study reveals that most people remain unwilling to trust AI, a fundamental privacy and adoption barrier.

of US adults say AI in daily life makes them more concerned than excited

Concern has risen steadily from 37% in 2021 to 50% in 2025. Only 10% say they are more excited than concerned.

of people would be less likely to engage with content marked as AI-generated

Only 18.51% would be more likely to engage. The AI label triggers privacy and trust concerns for the majority of users.

of internet users are at least a little worried about the environmental and data impact of AI

Over a third (34.46%) say AI's data and environmental footprint worries them 'a lot'. A ChatGPT query uses 10× the electricity of a Google search.

organisations have mature AI governance, the rest face elevated privacy and compliance risk

Only 20% of enterprises have established robust AI governance covering data privacy, model oversight, and shadow AI detection.

of organisations enforce a complete ban on generative AI, 97% must govern its data use

With near-universal AI usage, privacy policies must evolve from 'block AI' to 'secure AI data flows' across the entire organisation.

of people globally admit relying on AI output without evaluating its accuracy

Two-thirds of AI users don't verify AI responses, raising concerns about data accuracy, misinformation, and privacy of input data.

of search users have experienced at least one significant mistake in an AI Overview

42.1% encountered inaccurate content, 35.82% found missing context, and 16.78% received unsafe advice, privacy implications abound.

of enterprises rate their overall AI strategy, including privacy, as highly prepared

Less than half of organisations feel their AI privacy posture is strong, the remaining 58% acknowledge gaps.

average cost savings from extensive security AI, which includes data loss prevention

AI in security catches data exfiltration and privacy breaches faster, reducing costs by $2.22M per incident.

of organisations with AI functions have centralised risk and compliance, including data privacy

Risk and compliance is the most centralised AI function (57%), reflecting its importance for data privacy governance.

of organisations increased their AI investment since 2023, privacy infrastructure must scale to match

As AI investment surges, privacy controls must grow proportionally, yet many organisations are investing in capabilities before controls.

of Gen Z trust AI to be 'objective and accurate', vs. just 18% of Baby Boomers

A 31-point generational gap in AI trust shapes how privacy concerns are perceived. 45% of Boomers say 'I don't trust it' vs. 18% of Gen Z.

of women would be less likely to engage with AI-labelled content, vs. 42.54% of men

A significant gender gap in AI privacy concern means privacy messaging must be tailored to different demographics.

of people aged 60+ want the same amount or less AI-generated content online

Older adults are most resistant to AI content expansion, privacy and trust concerns grow stronger with age.

median share across 25 countries who are mainly excited about AI, the rest are concerned or mixed

In no country surveyed do more than 3 in 10 adults say they are mainly excited, privacy concern is globally pervasive.

more electricity used per ChatGPT request compared to a traditional Google search

AI's energy footprint raises privacy-adjacent concerns about data centre scale, water usage, and the sustainability of AI data processing.

of people say they always trust AI Overviews, meaning 91.5% have some privacy or accuracy scepticism

Only 8% always follow source links, yet most don't fully trust AI search, creating a disconnect between usage and privacy awareness.

of people globally say national and international AI regulation, including privacy rules, is needed

A strong public mandate for AI privacy governance across all 47 countries surveyed by KPMG and the University of Melbourne.

of people still believe AI will deliver benefits, privacy concerns haven't killed optimism entirely

Despite low trust (46%) and high concern, a strong majority see potential benefits, if privacy and governance can catch up.