5 ms·

Sadly we are well aware of how inaccurate and untrustworthy eye-witness accounts can be.

chrisseaton 6y ago
But at least there is a human being behind each statement. You can question, you can examine their motives, you can challenge. A video is just a video. An anonymous accuser. You can't ask a video any questions.
- ethbr0 6y ago
  Slightly OT, but I finally realized this was probably the long-game behind Facebook et al.'s real-identity-only policy. Sure it has the side (main!) benefit of empowering targeted advertising. But fundamentally creating traceability between content/comment and a singular identity enables elimination of the most egregious abuses (mass-disinformation/ganging). I'm as 90s internet-is-for-anonymous as anyone, but I have to admit traceability has merit in the larger social ecosystems.
  - intotheabyss 6y ago
    We can create decentralized identity systems, where you decide which information should be public, while also providing accountability. https://www.microsoft.com/en/security/business/identity/own-your-identity https://www.microsoft.com/en/security/business/identity/own-... Rather than let mega corps control identities, which should be building open source tools to allow developers to include decentralized identities within their platforms, preferably systems that interoperate with Ethereum.
  - pjc50 6y ago
    Much of the worst content is published by people under their own names and news bylines. And then happily retweeted by people under their own names. It gets worse when businesses and adverts are included in the mix. Who is Bob's News Agency, really? Can you click on every advert to find who paid for it? No.
    - ethbr0 6y ago
      Absolutely, but the difference is those people have fingerprints. If someone is repeatedly toxic (which should be algorithmically identifiable), you can take steps to balance that. You cannot do the same if 50% of your userbase lacks a stable / historical identity. At least without attempting to recreate identity on the basis of metadata (IP, patterns, etc).
- chillwaves 6y ago
  A video is data that exists with forensic context. What we need are institutions that are trusted to verify this data, like journalism. Except it is being allowed to be perverted for profit or ideology.
  - chrisseaton 6y ago
    What does a journalist know about verifying video data? And isn't a journalist going to bring their own side of the story to it? Hardly independent.
    - chillwaves 6y ago
      I'm sure the New York Times, with its resources, is able to hire some help. It's honestly incomprehensible to me that anyone would think otherwise. A journalist that has a reputation for truth would hopefully bring that integrity to the process. Every party will have some kind of perspective or bias, we should pick the ones who are incentivized towards truth vs lying for profit/sensation.
    - filoleg 6y ago
      >What does a journalist know about verifying video data? I don't think the parent comment was using journalists as a specific group of people who should be the ones in charge of verifying video data. I took it more as "we need a dedicated group of people for verifying video data, similarly to how we had journalists verifying all other sorts of data for their articles since the dawn of journalism". Goes without saying, but, also, the parent comment was clearly alluding to the "verification by journalists" in a traditional sense of proper investigative journalism, not "journalism" that is based on pulling random tweets without context from random no-names.
Apes 6y ago
As inaccurate and untrustworthy as deepfakes are making video?
- kube-system 6y ago
  If not worse. 1. Physical evidence can at least by analyzed by a third party, and 2. Misleading deepfakes aren't created accidentally by honest people.
asperous 6y ago
Eye witnesses are very accurate when it comes to broad ideas like did you hear a loud noise on this day. Where they fail, especially compared to evidence and video, is details like what did the noise sound like, what did the person look like, what were they wearing, etc

But at least there is a human being behind each statement. You can question, you can examine their motives, you can challenge. A video is just a video. An anonymous accuser. You can't ask a video any questions.

ethbr0 6y ago
Slightly OT, but I finally realized this was probably the long-game behind Facebook et al.'s real-identity-only policy. Sure it has the side (main!) benefit of empowering targeted advertising. But fundamentally creating traceability between content/comment and a singular identity enables elimination of the most egregious abuses (mass-disinformation/ganging). I'm as 90s internet-is-for-anonymous as anyone, but I have to admit traceability has merit in the larger social ecosystems.
- intotheabyss 6y ago
  We can create decentralized identity systems, where you decide which information should be public, while also providing accountability. https://www.microsoft.com/en/security/business/identity/own-your-identity https://www.microsoft.com/en/security/business/identity/own-... Rather than let mega corps control identities, which should be building open source tools to allow developers to include decentralized identities within their platforms, preferably systems that interoperate with Ethereum.
- pjc50 6y ago
  Much of the worst content is published by people under their own names and news bylines. And then happily retweeted by people under their own names. It gets worse when businesses and adverts are included in the mix. Who is Bob's News Agency, really? Can you click on every advert to find who paid for it? No.
  - ethbr0 6y ago
    Absolutely, but the difference is those people have fingerprints. If someone is repeatedly toxic (which should be algorithmically identifiable), you can take steps to balance that. You cannot do the same if 50% of your userbase lacks a stable / historical identity. At least without attempting to recreate identity on the basis of metadata (IP, patterns, etc).
chillwaves 6y ago
A video is data that exists with forensic context. What we need are institutions that are trusted to verify this data, like journalism. Except it is being allowed to be perverted for profit or ideology.
- chrisseaton 6y ago
  What does a journalist know about verifying video data? And isn't a journalist going to bring their own side of the story to it? Hardly independent.
  - chillwaves 6y ago
    I'm sure the New York Times, with its resources, is able to hire some help. It's honestly incomprehensible to me that anyone would think otherwise. A journalist that has a reputation for truth would hopefully bring that integrity to the process. Every party will have some kind of perspective or bias, we should pick the ones who are incentivized towards truth vs lying for profit/sensation.
  - filoleg 6y ago
    >What does a journalist know about verifying video data? I don't think the parent comment was using journalists as a specific group of people who should be the ones in charge of verifying video data. I took it more as "we need a dedicated group of people for verifying video data, similarly to how we had journalists verifying all other sorts of data for their articles since the dawn of journalism". Goes without saying, but, also, the parent comment was clearly alluding to the "verification by journalists" in a traditional sense of proper investigative journalism, not "journalism" that is based on pulling random tweets without context from random no-names.

Apes 6y ago

As inaccurate and untrustworthy as deepfakes are making video?

kube-system 6y ago
If not worse. 1. Physical evidence can at least by analyzed by a third party, and 2. Misleading deepfakes aren't created accidentally by honest people.

asperous 6y ago

Eye witnesses are very accurate when it comes to broad ideas like did you hear a loud noise on this day. Where they fail, especially compared to evidence and video, is details like what did the noise sound like, what did the person look like, what were they wearing, etc