No time for a proper post this week, but wanted to share a few thoughts about race and technology. This question, especially: what is the color of data?
In one sense, data has no color. Or maybe it’s the luminescent green of binary code 1’s and 0’s.
In another sense, data is white. As in the white folks who are working the jobs in data, big or otherwise. So maybe Asian too.
What I’m really after, though, is a visual of whose information is being shared in this era of information sharing. I’m interested in personal data – PII (personally identifiable information) – that, in the maw of the government, has personal consequences: arrest, harassment, deportation. This data, I suspect, is Black and Brown.
We can start with the basics, the decade-long effort to connect up the FBI’s Integrated Automated Fingerprint Identification System (IAFIS) and DHS’s Automated Biometric Identification (IDENT). These two databases contains, from IAFIS, the criminal records of 70 million people + noncriminal info on 34 million people + info on 73,000 known and suspected terrorists, and from IDENT, information on 55 million people. The racial disparities in the criminal justice system are cumulative, and so are most intense at the back end of the system (length of prison system). But even at the front end – where fingerprints are taken – the disparity is intense, particularly when policies like NYPD’s stop & frisk are in effect.
We can look as well at the information being sucked up and kicked around by HSIN, the Homeland Security Information Network. In its latest iteration, HSIN 3.0 now contains information about “members of the public who are the subject of documents, reports, or bulletins contained in the HSIN collaboration spaces.” Those collaboration spaces range far and wide, but the upshot is that these are largely reports on people whose “observed behavior may be indicative of intelligence gathering or pre-operational planning, related to terrorism, criminal, or other illicit intention.” The “see something say something” campaign has included some desultory efforts to encourage people to not find someone suspicious because of his race, but to report only suspicious behaviors.
We can look at intelligence information that’s being shared in the DHS Intelligence Enterprise (DHS IE), made up of the DHS Office of Intelligence & Analysis (I&A) + the intelligence elements of: U.S. Customs and Border Protection (CBP), U.S. Immigration and Customs Enforcement (ICE). U.S. Citizenship and Immigration Services (USCIS), the Transportation Security Administration (TSA), U.S. Coast Guard (USCG), and U.S. Secret Service (USSS).
There’s no question these data sets are disproportionately African American, Latino, South Asian, and Muslim. But how disproportionate? And why? And is that OK? That’s what I’d like some data on.