'Overseeing' the Poor: Technology Privacy Invasions of Vulnerable Groups

Scored by Grossman

Wendy Grossman has written an excellent article, Scored: Who's watching the watchers, that focuses on Birny Birnbaum's presentation on insurance and credit scoring that was part of Wednesday's plenary, Overseeing the Poor: Technology Privacy Invasions of Vulnerable Groups.

Issues with 'Credit Scoring' Presentation

Birny Birnbaum discussed the use of credit reports for purposes other than lending you money. Birnbaum suggests that we all know that banks are out to make money�He views this reality as being discriminatory against the low income and minority consumers and the young.

But how else do we propose that financial institutions manage risk? These factors have a track record for predicting the credit risk that individuals represents. Brinbaum equates these �biased� factors with de facto racial discrimination. It is true that these factors certainly present economic status discrimination and to the extent that economic status correlates with race and ethnicity, discrimination is indeed present, but this is a collateral effect.

It is unclear whether Birnbaum has any data to support a proposition that these factors are not predictive of risk�he suggests that financial responsible behavior would be a better measure. Accepting this proposition on this face, he does not suggest how to collect information on financial responsibility�would we need more information on consumers�how would that advance the goal of increasing privacy�would we not be advocating for even more intrusive data mining.

On the opposite side of spectrum, does Brinbaum suggest that we provide credit without assessing risk�should our society spread the cost of this risk�should the financially responsible subsidize the activities of the financially irresponsible?

If I am a safe driver, do I not feel entitled to paying less for my insurance, if I engage in risky behavior, should I not pay more for my insurance. It is acceded that the underlying premise of insurance on some level is risk spreading�maybe we need to revisit the concept of whether we should enforce non-differentiated insurance. But are we, the careful and risk averse, truly willing to subsidize the activities of the risk takers?

Insurance has long been recognized as the paradigmatic example of adverse selection�those who most fear the consequences of an accident or who are not judgment proof are the ones who most want insurance, they are also, arguably, the most careful.

Homeless and Privacy

PRIVACY FOR THE HOMELESS

Cindy Southworth, of the National network to End Domestic Violence, pointed out that there has been an initiative to collect information on the homeless to have a better understanding of the situation of the homeless. While this is an admirable goal, it requires the collection of data on victims of domestic violence who are fleeing their attackers�this collection of data threatens these women.

Southworth questions whether there is ever truly informed consent to the collection of this data and whether there are low tech, less invasive means of collecting reasonably accurate counts of homeless individuals (such as Thursday night shelter counts). She makes an impassioned plea to balance the value of detailed information on the homeless population against the cost of the lives of the victims of domestic violence who are traced through these central databases.

Chance Martin, of the SF Coalition on the Homeless, discussed HMIS. Collection of information on the homeless in HMIS systems are required by HUD�in order to be eligible for federal funding to assist the homeless you need to have such a system in place. These systems frequently require biometric finger scans and a digital portrait to identify the homeless.

There is both support and opposition among the homeless who are subjected to these systems. Some perceive it as a means for identifying criminals within the system and obtaining safer shelters. About an equal number object to such treatment as demeaning and treating all homeless as if they are criminals.

This is an unfunded mandate that promises little to help the homeless.

Chris Hoofnagle, of EPIC moderated the discussion. He pointed out that high technology is being used to monitor the technologically disenfranchisedis frightening new ways.

Partial Running Notes

Cindy Southworth: connection between homelessness database and domestic violence victims

HUD putting in place systems (like HMIS) that are almost as invasive as sex offender registries but instead target domestic violence victims and other support services users. With 1/3 women being battered in their lifetime, this database would be a massive clearinghouse of information on citizens. Massive tracking and surveilance system.

- fabulous tool for stalkers -- central server that is not very secure that contains identity and location information on victims who are already under threat.

There is some threat to shelters that if they don't give up their "victim data" they will lose funding.

info includes: medical health status, pregnancy status, marital status. All in exchange for being able to stay in a shelter for a night.

Info shared with 400 agencies including all staff. Insiders who are abusers will have a field day.

Informed consent is a joke when working with victims of imminent violence.

Chance Martin: The CHANCE System

All of the biometric database systems cost millions of dollars (not funded by the Feds) and haven't really helped crack down on fraud or abuse in the system. There are also problems technologically with the systems.

Survey: 30% said they would rather stay on the streets than submit to biometric tagging at shelters. 50% among latinos, which is probably related to immigration issues. Many says #1 concern is invasion of privacy.

If enough political pressure is brought to bear in the SF political market, it might be possible to defund the HMIS/biometric effort.