- From: Katryna Dow <katryna.dow@meeco.me>
- Date: Sat, 16 Jun 2018 03:50:20 +0000
- To: Moses Ma <moses.ma@futurelabconsulting.com>
- CC: "W3C Credentials CG (Public List)" <public-credentials@w3.org>
- Message-ID: <8F6ACB9B-B3EF-4EB0-AB88-1EE7FD638919@meeco.me>
Hi Moses, I’m the Chair of IEEE’s #P7006 - the standard for a single (human) AI is exactly what we’re working on. One of the aims is for the AI to be able to process T&Cs or “offers” and assess bias and downstream consequences. If your interested in contributing to the work, let me know. Katryna [Meeco]<http://www.meeco.me/> Katryna Dow CEO & Founder, Meeco Australia: +61 418 538 404<tel:+61%20418%20538%20404> USA: +1 201 942 0009<tel:+1%20201%20942%200009> UK: +44 7408 854 221<tel:+44%207408%20854%20221> katryna.dow@meeco.me<mailto:katryna.dow@meeco.me> Twitter: @katrynadow<https://twitter.com/katrynadow> LinkedIn: katrynadow<http://www.linkedin.com/in/katrynadow> Skype: katryna.dow<skype:katryna.dow?userinfo> Discover Meeco at meeco.me<http://www.meeco.me/> Follow us on twitter @meeco_me<https://twitter.com/meeco_me> and blog.meeco.me<http://blog.meeco.me/> On 16 Jun 2018, at 12:04, Moses Ma <moses.ma@futurelabconsulting.com<mailto:moses.ma@futurelabconsulting.com>> wrote: Hi everyone, I just wrote an article about psychometrics — the stuff Cambridge Analytica used — and how it might interact with decentralized ID. See: https://www.psychologytoday.com/ca/blog/the-tao-innovation/201806/the-psychometrics-decentralization I'm wondering if there might be some way to harden the DID system against manipulative psychometrics. I mean, beyond non-correlation. In the article, I proposed a fanciful intelligent assistant app, like a Siri for the forces of good, that knows your psychometric type, and whenever something looked manipulative in an ad or in your Facebook feed, it could check to see the reputation of the poster and warn you. However, this would require a framework for reputation. Instead, I'm thinking there may be something easier and simpler. For example, instead of requiring some proof of work – via Equihash - to post something, how about using Equihash to make it through your spam filter? Or better yet, send a few cents to a charitable cause, to make it into your feed, ad queue, inbox. You have any ideas? Moses PS, if you could retweet these articles, I'd appreciate it: * https://twitter.com/mosesma/status/1007328677892272128 * https://twitter.com/mosesma/status/1007289955486322688
Received on Saturday, 16 June 2018 03:53:56 UTC