Big Data for Good or Evil -- Lessons from the NSA PRISM Scandal
summary
Argues that PRISM shows organizations must govern not just desired data but call detail records and other metadata that reveal identity; claims NSA analytics reduce global traffic to 7.5 terabytes per day — roughly 5 million phone calls — exposing how small extracts can still be massive. Advises practical tools and risks: use Hadoop for mixed-quality data, adopt WS-Policy style metadata for automated enforcement, beware false positives and misconfiguration (e.g. studies finding 70% of S3 buckets unsecured), and build meta‑policies to keep analytics from creating harmful or illegal exposures.
AI-written summary · about
chapters (6)
- 0:00Start
- 1:402012 Big Data Technology Landscape
- 5:03Big Data Crisis Point
- 6:06Parkinson's Law (Big Data Corollary)
- 21:02Today's Data Governance (simplified)
- 24:41Dealing with Change
related talks
01h 16mData protection and Big Data: what about regulating Big Data analytics?
24mData’s Rotten Scoundrels – How To Avoid Becoming One
01h 04mData Hemorrhage, Inequality, and You...
20mEconomic Impact of PRISM on Cloud Services & Safe Harbor
47mDesigning for User-Centered Privacy
01h 27mTrans-atlantic discussions: Developing a sustainable 'Big Data' ecosystem ...
14m"Big Data" and Analytics in Government
01h 14mEthical data handling