Jennifer Cobbe Profile picture
Tech law @Cambridge_CL
Apr 28, 2023 5 tweets 2 min read
Kenneth Baker warning of the dangers of the digital welfare state when introducing his Data Surveillance Bill in *1969* Image Source: api.parliament.uk/historic-hansa…
Apr 28, 2023 4 tweets 1 min read
Looking for the earliest use of the term "personal data" to mean roughly what it means today (information about a person). The earliest I've found is 1911 (in a US government report on Japanese immigrant communities). Any advance on that? A brief history of "personal data" Image
Apr 28, 2023 18 tweets 4 min read
FAccT paper with @mikarv and @jatinternet - on accountability in algorithmic supply chains

We explore the dynamics of AI supply chains, their implications for accountability, and why we need to understand their legal and political economic context

Link: papers.ssrn.com/sol3/papers.cf… A lot of algorithmic accountability work is based (implicitly or otherwise) in trying to understand the actions of a single organisation which makes and deploys an AI system

These tools are useful but this focus is a mistake
Jul 18, 2022 30 tweets 5 min read
Skimming this new data protection Bill, one thing for sure is it's going to make the Data Protection Act 2018 and the UK GDPR even more of a mess to follow What is the point of abolishing the Information Commissioner (and his office) and replacing him with an Information Commission with a chair and the same role and responsibilities? Other than to waste money on new headed paper
Dec 21, 2021 13 tweets 3 min read
The reason why GDPR hasn't made a difference here is that data protection law - as essentially constituted since at least the 1990s - is itself part of the problem

For GDPR to have changed anything it needed a much more radical departure from the core framework of the Directive Law is part of what produces various social, political economic, and technological processes and practices, including tracking

Tracking emerged in the context of law that permitted it, through its framework, its concepts, its approach, and its effects
Dec 21, 2021 5 tweets 2 min read
Women & Equalities Committee calls on the government to move towards a self-ID model for trans people They also recommend that the requirement to 'live in the acquired gender' for two years - which perpetuates outdated stereotypes - be removed *immediately*
Oct 6, 2021 20 tweets 6 min read
Seven more thoughts on recommender systems

ONE. Law needs to identify specific platform activities to regulate, rather than kinds of technology

We identify activity of 'recommending' - constructing content feeds algorithmically based on determination of relevance, interest, etc But 'recommending' is itself too broad a description of what platforms are doing. So we divide 'recommending' into three types: open, curated, and closed

The distinction between these activities turns on the platform's sourcing of content for the recommender system
Oct 6, 2021 9 tweets 2 min read
In 2019 we proposed that chronological should be the default, with algorithmic opt-in only, and controls available for users who choose it

Also: specific transparency requirements for platforms around recommenders for user content

Read all about it: papers.ssrn.com/sol3/papers.cf… Opting-in and controls for users is ofc limited in that platforms will manipulate this (as they do with, e.g., data protection controls)

So we also argue for prohibiting certain platforms from using recommenders for user content when they've shown they can't do so responsibly
Jun 7, 2021 12 tweets 5 min read
New paper coming in @CompLawSecRev - "Artificial Intelligence as a Service: Legal Responsibilities, Liabilities, and Policy Challenges"

We examine the roles, responsibilities, and power of AI as a Service providers

Link: papers.ssrn.com/sol3/papers.cf… We talk about what AIaaS is and how it differs from other cloud services, and set out the chain of processing as personal data flows from data subjects to customers to providers and back again
Mar 28, 2021 4 tweets 1 min read
This is all sorts of bad

The pandemic must not be used to normalise facial recognition as a condition for access to everyday things If this goes ahead, it will end up being used to check for other things beyond covid status. Police, immigration, welfare agencies will be all over it

And once it's normalised for going to the pub, why not shops? Why not offices? Why not transport? Why not everywhere else?
Oct 8, 2020 13 tweets 3 min read
My paper on algorithmic censorship now published by Philosophy & Technology - I assess what auto content moderation means for power of social platforms and their ability to put commercial priorities further into everyday communications of billions of users
link.springer.com/article/10.100… An underappreciated thing is automated content moderation can succeed as a project of power (sanitising platforms, making them more commercially acceptable, disciplining users) even if fails as a project of moderation (the systems are flawed and some users will get around them)
Jun 22, 2020 6 tweets 1 min read
A company builds a computer system - with various features, limitations, restrictions, and so on, as with any computer system - and it becomes widely used

Later, a government wants to do something with that computer system that isn't supported technically Without working any of the legal or regulatory levers available to it, the government simply asks the company to allow them to do the thing they want to do

The company instead implements a mechanism that would allow for a similar but functionally quite different thing
Dec 5, 2018 4 tweets 1 min read
Another 'ditch Facebook' think piece which ignores the reality that Facebook has actively embedded itself in more than just social lives. It just isn't that simple for many people. Unless and until you understand that fact you're miles from a workable solution to Facebook's harms 'Ditch Facebook' proposes an individual response to problems far beyond individual control. The issues with Facebook arise primarily and jointly from surveillance capitalism and the monopoly power of dominant platforms. Why individualise for solutions to a systemic disease?
Apr 16, 2018 6 tweets 3 min read
Facebook is now explicitly asking European users for consent to facial recognition This is the screen where they ask for consent. No clear option for 'no'

cc @darkpatterns