Dubai Police: Deepfakes Increasingly Hard to Detect

What happened

Dubai Police has warned that deepfake videos, images and voice recordings are becoming increasingly difficult to distinguish from genuine content, as artificial intelligence tools used to create them continue to advance. The warning stems from a study by the force's Future Foresight Centre, which found that rapidly improving deepfake technologies pose growing risks to public safety, the credibility of information, and the reliability of digital evidence used in decision-making.

The research found that the challenge goes beyond simply spotting manipulated material: as synthetic content becomes more convincing, it also erodes general confidence in what people see and hear online. The study warns that this could make it easier for bad actors to impersonate individuals, commit identity theft, mislead the public and spread disinformation.

The findings were released to coincide with Cybersecurity Awareness Month, marked every October, and form part of the UAE Ministry of Interior's "Safe UAE Through Your Digital Awareness" campaign.

Why it matters

The study underscores how generative AI is reshaping the baseline assumptions institutions, businesses and the public rely on when assessing authenticity. As deepfakes become harder to detect with the naked eye or ear, organisations that depend on video, audio or image verification — from banks confirming customer identity to courts assessing evidence — face a growing credibility gap that technology alone may struggle to close.

For digital transformation and public-sector leaders, this signals a shift from treating synthetic media as a niche security concern to recognising it as a mainstream operational risk. Verification processes, fraud controls and digital evidence handling may all need to be reassessed as the line between authentic and fabricated content continues to blur.

The Renascence take

Most coverage of deepfakes focuses on the technology race between detection tools and generation tools. The more interesting shift is behavioural: once people learn that convincing fakes exist, they start doubting real content too, not just fake content.

The real cost of advancing deepfake technology isn't that people will believe fake things — it's that they'll stop trusting real ones. That erosion of default trust changes how customers behave at every verification touchpoint, from a call centre voice check to a video submitted as proof of identity. Organisations that rely on "seeing is believing" verification need to start designing for a world where authenticity has to be actively earned and demonstrated, not assumed. The operators who get ahead of this will treat verification friction as a trust-building feature, not a nuisance to be minimised.