18 Matching Annotations
  1. Aug 2023
    1. DIGITAL

      Scope seems confusing because whenever the term 'processing' is used, it just applies to digital personal data, but if the term processing is not used, it applies to all personal data?

    2. wholly or partly automatedoperation or set of operations performed on digital personal data

      Processing narrower in scope than EU law because only applies to digital personal data, not records in a filing system.

    3. (5) The Central Government may, before expiry of five years from the date ofcommencement of this Act, by notification, declare that any provision of this Act shall notapply to such Data Fiduciary or classes of Data Fiduciaries for such period as may bespecified in the notification

      Another big potential loophole that firms can lobby for no?

    4. ensure its completeness, accuracy and consistency.

      Accuracy principle only applies in the case data is used for decisions or disclosed. Not a passive duty for all datasets processed.

    5. erase

      Does not appear to be a right to object in this law, which is in many ways stronger than erasure as you object to a purpose, while erasure fails if a purpose is still valid.

    6. he identities of all other Data Fiduciaries and Data Processors with whomthe personal data has been shared by such Data Fiduciary, along with a description ofthe personal data so shared

      This is stronger than the text of the GDPR in relation to recipients, which mandates only 'categories' to be provided, although the CJEU has strengthened this in a recent case.

    7. summary of personal data

      Not a copy as per art 15, GDPR. No clear guidance on what constitutes an adequate summary or not. Would not apply to e.g. emails, text and so on very well, which access requests are widely used for in other jurisdictions as accountability tools.

    8. specific to a Data Principal

      But potentially used by a downstream actor, as is the case with credit scoring. Would a decision taken by a bank on behalf of a credit scoring agency be 'specific' to the credit scoring agency?

  2. Jan 2018
    1. anonymous

      Core question: are the identifiers persistent at all? If not, then trivially easy to triangulate data to work out who-is-who.

  3. Dec 2017
    1. Can an employer demand that you go to work naked?

      this is a very misleading article that does not mention or discuss anythign in its title...

  4. Feb 2017
    1. If a private university such as BPP can provide quality law degrees for £6,000 a year, why can’t others? It’s likely that some universities are using fees to cross-subsidise research. In others, fees will help pay for flabby management, overly generous leadership salaries and big capital investment projects, even though, in surveys, that’s not what students say they prioritise.

      Students consistently say that the space they learn and study in is an important thing. Many universities are justified in improving their facilities in a forward looking way using fees.

    2. There’s a dinosaur-like tendency to do things the way they’ve always been done.

      Weren't dinosaur's innovative?

  5. May 2016
    1. “equivalent” in assessing defendants of different races.

      ML scholars generally believe accuracy is a poor metric, especially if your data is skewed.

    2. The trick, of course, is to make sure the computer gets it right. If it’s wrong in one direction, a dangerous criminal could go free. If it’s wrong in another direction, it could result in someone unfairly receiving a harsher sentence or waiting longer for parole than is appropriate.

      Actually the trick is surely to weigh up whether false positives or false negatives are more important to you

    3. derived

      How? Through a learning machine or through manual point allocations?

    4. the algorithm was somewhat more accurate than a coin flip

      In machine learning it's also important to evaluate not just against random, but against how well other methods (e.g. parole boards) do. That kind of analysis would be nice to see.

  6. Nov 2015
    1. that a popular local market had been cancelled. In fact, it was operating as normal.

      quite a struggle to govern informal systems, as many have found...