[Air-L] New CDT research report - "Do You See What I See? The Capabilities and Limitations of Automated Multimedia Content Analysis"

[Dhanaraj Thakur]

Hi everyone,

We published a new report today that explains how machine learning 
techniques are used to assess user-generated content online such as 
images, video, and audio media. It also explains what these automated 
tools can—and cannot—tell us about multimedia content. The report, /Do 
You See What I See? The Capabilities and Limitations of Automated 
Multimedia Content Analysis <https://cdt.org/do-you-see-what-i-see>/, is 
accompanied by an infographic 
<https://cdt.org/wp-content/uploads/2021/05/CDT-Five-Limits-of-Automated-Multimedia-Content-Analysis.pdf> 
that summarizes five main limitations of using these automated tools at 
scale. We argue that researchers, civil society advocates, policy-makers 
and others should be aware of these limitations particularly when trying 
to address content moderation issues.

We also have a twitter thread 
<https://twitter.com/CenDemTech/status/1395363930638364674> with a few 
highlights from the report.

Feel free to share, and it would also be great to get your feedback.

thanks,

Dhanaraj


-- 

*Dhanaraj Thakur* (he/him) | Research Director
Center for Democracy & Technology |*cdt.org <https://cdt.org/>*
*E:* dthakur at cdt.org <mailto:dthakur at cdt.org> | *P:* +1 202 407 8849 | 
@thakurdhanaraj