5 ms·

The question is, how?

by jiggy2011 12y ago

1123581321 12y ago
There are two methods, an old one and a new one. The old one is to spend more time with fewer people. The new one is to know more about new people as soon as you meet them. This is one of the goals of projects like Google Glass.
- jiggy2011 12y ago
  How would google glass know whether somebody was trustworthy or not?
  - 1123581321 12y ago
    Imagine talking to someone at a conference and seeing a summary of their expertise from StackOverflow, an estimate of their credit score or their peer reviews from GlassDoor or LinkedIn right next to their head. Imagine talking to someone and using an app that assigns a probability the person is telling the truth based on language, tone, facial changes and reviews from other users of that app. Those kinds of things are part of Google's ambition for glass. Even now, Google can tell you someone's name and where they're from, allowing you to catch basic lies you wouldn't normally notice. Edit: Given the downvotes without explanation, I have to explain that I'm not a fan of Google glass - I'm just explaining that Google wants to instantly deliver information about people that currently is learned over time.
    - Infinitesimus 12y ago
      >Imagine talking to someone and using an app that assigns a probability the person is telling the truth based on language, tone, facial changes and reviews from other users of that app. That would be no better than our current biological means of handling the problem though. They both rely on analyzing data from sources that may or may not be accurate and assigning probabilities. What happens to the person who is given a high probability of being a liar because we have biased data based on race, sex, gender etc? The problem is indeed very complicated and I'm not convinced technology will do a better job at that. Tech doesn't know that this person who has a poor review on linkedIn or whatever was actually suffering through loss of a close friend, relationship issues, etc. which all affected his/her mood. As much as I love offloading tasks to automated systems, I fear that we might go too far and offload some of the things that make up our core humanity too.
      - 1123581321 12y ago
        I agree with your overall opinion about humanity, and I am too familiar with credit reports to think they are perfect arbiters. However, we have to ask what exactly we are wanting to trust when we meet people. Once we know what that is, then we can develop technology to receive the information earlier in the relationship. This already happens, just not in time to affect the first impression. When we meet people, we don't know anything about them, so if we have the goal of removing bias from first encounters with people, we either have to train people assume nothing about them, which is foolish and probably impossible, or provide (hopefully unbiased) information that is valued more than what the person perceives to be the benefit of the bias. (Again, I'm just speculating about this.)
    - Nursie 12y ago
      That sounds awful! And a complete invasion of privacy. I really don't want to live in that world.
  - nl 12y ago
    With something like IARPA INSTINCT[1]. It was won by JEDI MIND[1] (capitals because it's an abbreviation, which deserves credit in itself). [1] http://www.iarpa.gov/index.php/working-with-iarpa/prize-challenges/218-instinct-iarpa-s-trustworthiness-challenge http://www.iarpa.gov/index.php/working-with-iarpa/prize-chal... [2] http://www.nextgov.com/emerging-tech/2014/10/force-strong-one-iarpa-picks-jedi-mind-predict-intelligence-trustworthiness/96191/ http://www.nextgov.com/emerging-tech/2014/10/force-strong-on...

The question is, how?

There are two methods, an old one and a new one. The old one is to spend more time with fewer people. The new one is to know more about new people as soon as you meet them. This is one of the goals of projects like Google Glass.

jiggy2011 12y ago
How would google glass know whether somebody was trustworthy or not?
- 1123581321 12y ago
  Imagine talking to someone at a conference and seeing a summary of their expertise from StackOverflow, an estimate of their credit score or their peer reviews from GlassDoor or LinkedIn right next to their head. Imagine talking to someone and using an app that assigns a probability the person is telling the truth based on language, tone, facial changes and reviews from other users of that app. Those kinds of things are part of Google's ambition for glass. Even now, Google can tell you someone's name and where they're from, allowing you to catch basic lies you wouldn't normally notice. Edit: Given the downvotes without explanation, I have to explain that I'm not a fan of Google glass - I'm just explaining that Google wants to instantly deliver information about people that currently is learned over time.
  - Infinitesimus 12y ago
    >Imagine talking to someone and using an app that assigns a probability the person is telling the truth based on language, tone, facial changes and reviews from other users of that app. That would be no better than our current biological means of handling the problem though. They both rely on analyzing data from sources that may or may not be accurate and assigning probabilities. What happens to the person who is given a high probability of being a liar because we have biased data based on race, sex, gender etc? The problem is indeed very complicated and I'm not convinced technology will do a better job at that. Tech doesn't know that this person who has a poor review on linkedIn or whatever was actually suffering through loss of a close friend, relationship issues, etc. which all affected his/her mood. As much as I love offloading tasks to automated systems, I fear that we might go too far and offload some of the things that make up our core humanity too.
    - 1123581321 12y ago
      I agree with your overall opinion about humanity, and I am too familiar with credit reports to think they are perfect arbiters. However, we have to ask what exactly we are wanting to trust when we meet people. Once we know what that is, then we can develop technology to receive the information earlier in the relationship. This already happens, just not in time to affect the first impression. When we meet people, we don't know anything about them, so if we have the goal of removing bias from first encounters with people, we either have to train people assume nothing about them, which is foolish and probably impossible, or provide (hopefully unbiased) information that is valued more than what the person perceives to be the benefit of the bias. (Again, I'm just speculating about this.)
  - Nursie 12y ago
    That sounds awful! And a complete invasion of privacy. I really don't want to live in that world.
- nl 12y ago
  With something like IARPA INSTINCT[1]. It was won by JEDI MIND[1] (capitals because it's an abbreviation, which deserves credit in itself). [1] http://www.iarpa.gov/index.php/working-with-iarpa/prize-challenges/218-instinct-iarpa-s-trustworthiness-challenge http://www.iarpa.gov/index.php/working-with-iarpa/prize-chal... [2] http://www.nextgov.com/emerging-tech/2014/10/force-strong-one-iarpa-picks-jedi-mind-predict-intelligence-trustworthiness/96191/ http://www.nextgov.com/emerging-tech/2014/10/force-strong-on...