4 ms·

How people 'of no security interest' are keystrokes away in UK's spy databases

65 points by mocko 10y ago

deepnet 10y ago
collateral intrusion is a wonderful neologism.
- confluence 10y ago
  My favourite is parallel construction.
jsingleton 10y ago
Credit to Privacy International for unearthing all this: https://privacyinternational.org/node/843 https://privacyinternational.org/node/843 If you want to help this sort of work consider donating to: https://www.dontspyonus.org.uk https://www.dontspyonus.org.uk
Natanael_L 10y ago
Yet another reason to improve cryptography and get everything to use it
- schoen 10y ago
  I'm not sure that all of the databases described in this article were obtained via signals intelligence. Particularly because the article describes legislative authority to demand information from British companies. Ubiquitous encryption is a great solution to wiretapping, but nothing that we can unilaterally deploy protects most of the other aspects of our privacy. One thing that would clearly help is being able to do more things anonymously, but that ability is often limited by laws or companies' business practices that crypto alone won't solve. Sometimes crypto offers a way to do something anonymously, but the counterparty needs to be willing to participate in the privacy-protective protocol (for example, privacy-preserving metering and tolling schemes). If the counterparty says "I'm just going to record your fine-grained usage data with a traditional meter" or "I'm just going to record with RFID whenever you cross the toll bridge", the fact that an alternative has been invented doesn't help you. If we invent a privacy-preserving presence mechanism that could allow mobile phones to be anonymous except when actively making a phone call, or a way to pay for mobile data anonymously without revealing the identity of the SIM or device, we still don't get a privacy win unless mobile companies are willing to deploy it on their networks. If ZeroCash is deployed, we still don't get a privacy win in any particular area of the economy unless sellers are actually willing to accept it. Privacy advocates started worrying a lot about databases in the 1970s as database and storage technology improved. Privacy writers called them "databanks" back then, and they were probably the major privacy preoccupation of the day. All of these concerns sound super-quaint now that big data is so ubiquitous, but at the same time the concerns were largely correct! And the databank-preoccupied privacy advocates weren't necessarily thinking that the databanks would be filled up with data from wiretaps or spying on the Internet (which most people had never heard of at the time). They envisioned people stockpiling or obtaining a lot of transactional records, and making people register in order to do things, and linking one database to another by joining on common fields. And that's happened.
rlpb 10y ago
A search should automatically trigger a notification to the subject of the search in 12 months. Active investigations could maintain a list of people to whom notifications will be postponed. Investigators wanting to postpone notification by over 12 (further) months should be required to go to a judge for an extension. In all cases, the default position should be that the subject of a search will be notified within 12 months.

65 points by mocko 10y ago

deepnet 10y ago

collateral intrusion is a wonderful neologism.

confluence 10y ago
My favourite is parallel construction.

jsingleton 10y ago

Credit to Privacy International for unearthing all this: https://privacyinternational.org/node/843 https://privacyinternational.org/node/843 If you want to help this sort of work consider donating to: https://www.dontspyonus.org.uk https://www.dontspyonus.org.uk

Natanael_L 10y ago

Yet another reason to improve cryptography and get everything to use it

schoen 10y ago
I'm not sure that all of the databases described in this article were obtained via signals intelligence. Particularly because the article describes legislative authority to demand information from British companies. Ubiquitous encryption is a great solution to wiretapping, but nothing that we can unilaterally deploy protects most of the other aspects of our privacy. One thing that would clearly help is being able to do more things anonymously, but that ability is often limited by laws or companies' business practices that crypto alone won't solve. Sometimes crypto offers a way to do something anonymously, but the counterparty needs to be willing to participate in the privacy-protective protocol (for example, privacy-preserving metering and tolling schemes). If the counterparty says "I'm just going to record your fine-grained usage data with a traditional meter" or "I'm just going to record with RFID whenever you cross the toll bridge", the fact that an alternative has been invented doesn't help you. If we invent a privacy-preserving presence mechanism that could allow mobile phones to be anonymous except when actively making a phone call, or a way to pay for mobile data anonymously without revealing the identity of the SIM or device, we still don't get a privacy win unless mobile companies are willing to deploy it on their networks. If ZeroCash is deployed, we still don't get a privacy win in any particular area of the economy unless sellers are actually willing to accept it. Privacy advocates started worrying a lot about databases in the 1970s as database and storage technology improved. Privacy writers called them "databanks" back then, and they were probably the major privacy preoccupation of the day. All of these concerns sound super-quaint now that big data is so ubiquitous, but at the same time the concerns were largely correct! And the databank-preoccupied privacy advocates weren't necessarily thinking that the databanks would be filled up with data from wiretaps or spying on the Internet (which most people had never heard of at the time). They envisioned people stockpiling or obtaining a lot of transactional records, and making people register in order to do things, and linking one database to another by joining on common fields. And that's happened.

rlpb 10y ago

A search should automatically trigger a notification to the subject of the search in 12 months. Active investigations could maintain a list of people to whom notifications will be postponed. Investigators wanting to postpone notification by over 12 (further) months should be required to go to a judge for an extension. In all cases, the default position should be that the subject of a search will be notified within 12 months.