Pangram has expanded its AI detection tools to cover images and videos, extending a system that flags machine-generated content beyond text into visual media.
AI detection · Pangram
Pangram extends AI detection from text to images and video
Users can now check whether pictures and video clips were likely produced or manipulated by AI. Independent accuracy data for the new visual checks has not yet been published, so the reliability of the verdicts remains an open question.
How a visual check fits into verification
A single detector verdict is one signal, not proof of origin.
STEP 1
Image or video submitted
→
STEP 2
Detector estimates likelihood of AI origin
→
STEP 3
Verdict weighed with other evidence
Why it matters
Realistic AI images and video are easier to make than ever.
Newsrooms, platforms, educators and businesses need to verify media before publishing or acting on it.
Visual formats now sit inside the same verification workflow as text.
What remains unclear
No independent benchmarks or published error rates for image and video checks.
Authentic material can be wrongly flagged as synthetic.
Newer generation models may evade detectors not trained on them.
Indicators to watch
Performance data will decide how far these tools can be trusted.
Third-party evaluations
Whether Pangram publishes independent tests of the image and video detectors.
Edited or compressed media
How the tools perform on heavily altered or compressed files.
Competing vendors
Whether other detection vendors add similar multimodal coverage.
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article...