Published • loading... • Updated
Building Tool to Help Courts Spot AI-Generated Evidence Is ‘Challenging’: Researcher
Researchers say the system is still struggling to reach consensus in real-world tests, underscoring the limits of AI detection for courts.
- On Friday, Maura Grossman, co-director of a Canadian research team, noted that developing a reliable tool to detect AI-generated evidence for courts is proving more challenging than expected.
- Launched last year under the Canadian Institute for Advanced Research, the project aims to create open-source software because existing commercial tools lacked the transparency and reliability required for judicial contexts.
- Technical barriers like image compression and metadata loss complicate detection, Grossman said, and algorithms failed to reach consensus during two real-life test cases.
- Recognizing that a single tool is insufficient, researchers are now adopting a more modest approach that requires broader policy changes, regulation, and public education.
- The team plans to incorporate input from law enforcement and policymakers over the next year to address what Grossman describes as a "wicked technical problem.
Insights by Ground AI
24 Articles
24 Articles
The Toronto Star+4 Reposted by 4 other sources
Building tool to help courts spot AI-generated evidence is ‘challenging’: researcher
A researcher working to help courts detect evidence that's been created or altered with artificial intelligence says it's proving difficult to come up with a tool that's as effective in
·Toronto, Canada
Read Full Article
Winnipeg Free Press+18 Reposted by 18 other sources
Building tool to help courts spot AI-generated evidence is 'challenging': researcher
A researcher working to help courts detect evidence that's been created or altered with artificial intelligence says it's proving difficult to come up with a tool that's as effective in the real world as it seems to be in the lab.
·Winnipeg, Canada
Read Full ArticleCoverage Details
Total News Sources24
Leaning Left16Leaning Right0Center7Last UpdatedBias Distribution70% Left
Bias Distribution
- 70% of the sources lean Left
70% Left
L 70%
C 30%
Factuality
To view factuality data please Upgrade to Premium
