3 ms·

I have zero interest in assisting the government spy on people. Some spying, e.g. during wartime such as the work done at Bletchley Park, may be legitimate, but

I have zero interest in assisting the government spy on people. Some spying, e.g. during wartime such as the work done at Bletchley Park, may be legitimate, but given our current ongoing state of emergency and war against a non-state enemy, it is far more prone to abuse—as we've been seeing lately.

relaunched 13y ago
Unfortunately, the nature of classication is one where you need training data from both groups, those who are and are not people of interest. However, if you could use your skills to make sure we have less false positives, would you? Spying is contentious and that's part of the reason I asked this question. The other part is, whether we choose to believe it or not, we might be able to keep the country better than randomly safe if someone statistically attempts to classify those who intend to do us harm vs. those who don't. And, like it or not, the amount of info that's required to attempt to do this well is large, personal and intrusive. I don't disagree that someone could potentially SELECT * FROM DB WHERE NAME='CANDIDATE JOHN SMITH' AND MESSAGE_TO=(FEMALE != WIFE) AND COUNT >= 5 But, on the other hand, it could also find sleeper cells. What say you?
- LoganCale 13y ago
  The ends do not justify the means. One price of living in a free society that does not spy on its citizens is the greater risk of attack, and I find that an acceptable tradeoff. Total security is an impossible goal, even if you destroy the right to privacy in its pursuit.
- DanBC 13y ago
  I tend to agree. I think a government organisation with strict oversight who gathers enormous database of content is less worrying than various other groups of people who collect similar data. I'm more worried about employees at my cell phone company, for example. Facebook have a lot of information, but employees tell us that it's just too boring to go looking through people's profiles. (Also, they have audit trails and people do get sacked for unauthorised access.) While they shouldn't have content data, and they mustn't look at that content data without a warrant, I just don't care. I care even less about traffic 'meta' data.

relaunched 13y ago

Unfortunately, the nature of classication is one where you need training data from both groups, those who are and are not people of interest. However, if you could use your skills to make sure we have less false positives, would you? Spying is contentious and that's part of the reason I asked this question. The other part is, whether we choose to believe it or not, we might be able to keep the country better than randomly safe if someone statistically attempts to classify those who intend to do us harm vs. those who don't. And, like it or not, the amount of info that's required to attempt to do this well is large, personal and intrusive. I don't disagree that someone could potentially SELECT * FROM DB WHERE NAME='CANDIDATE JOHN SMITH' AND MESSAGE_TO=(FEMALE != WIFE) AND COUNT >= 5 But, on the other hand, it could also find sleeper cells. What say you?

LoganCale 13y ago
The ends do not justify the means. One price of living in a free society that does not spy on its citizens is the greater risk of attack, and I find that an acceptable tradeoff. Total security is an impossible goal, even if you destroy the right to privacy in its pursuit.
DanBC 13y ago
I tend to agree. I think a government organisation with strict oversight who gathers enormous database of content is less worrying than various other groups of people who collect similar data. I'm more worried about employees at my cell phone company, for example. Facebook have a lot of information, but employees tell us that it's just too boring to go looking through people's profiles. (Also, they have audit trails and people do get sacked for unauthorised access.) While they shouldn't have content data, and they mustn't look at that content data without a warrant, I just don't care. I care even less about traffic 'meta' data.