Special mention to our friends at @TBIJ who obtained key emails at the start of this case, showing Palantir wooing NHS execs and UK officials over Davos chats and watermelon cocktails.
And hats off to our amazing barristers, who worked so, so hard on this - Victoria Wakefield QC of @brickcourt, Julianne Kerr Morrison of @moncktonlaw, and Jennifer MacLeod, also at Brick. 🥰
The future of the NHS matters. The NHS is sitting atop a massive £10bn/a year trove of health data. US tech firms like Palantir hope to access or manage it for the long term.
Hancock’s plans for a revamp could build a future-fit NHS, or clear the path for US tech cos to profit.
We need an NHS we can trust. That means no access to data people don’t know about and agree to (the datasets fed into the datastore are STILL redacted).
It means no shady business partners (Palantir’s main clients are the CIA, cops, US border forces, and now the Home Office).
And it means no deals like this without getting our permission.
Palantir snuck in via emergency COVID-related contracts - now they’re trying to weasel their way into the future of our nation’s healthcare system.
Palantir fails the trust test. Their founder and chair is far-right Trump donor Peter Thiel, whose stated agenda is to extend his company’s reach across government.
2/ We love @signalapp. (We used to use @WhatsApp lots until Facebook took our data!)
Disappearing messages are great for us, the citizens. They're not appropriate for officials.
Why, you ask?
3/ Simple. The Public Records Act 1958 requires officials to review every message about the formulation of government policy to perform a legal check – in case it needs to be archived for public release.
If the message explodes, the check can never happen.
Last year, the government signed the largest data deal in history between the NHS and giant tech firms – like Palantir and Faculty, a British AI start-up involved with the Vote Leave campaign.
We took legal action and, with your support, forced them to publish the contracts.
Breaking: More than 200 @Facebook content moderators (and supporting employees) from across the US and Europe have published a demand for better protections - and full employment rights - during the pandemic.
This is the biggest joint international effort of Facebook content moderators yet. We are proud to have worked with them to do it. Many more moderators in other sites wanted to sign, but were too intimidated by Facebook - these people are risking their livelihood to speak out.
They've taken that risk because Facebook has risked their lives by forcing content moderators back to the offices. Full FB staff are allowed to work from home - many mods have to come in, even with live Covid cases on the floor, as @samfbiddle reported: theintercept.com/2020/10/20/fac…
NEW: last night, @Ofqual pulled from its website its guidance about the algorithmic grading appeals process. We’re posting it and a letter we've sent this AM because we think it opens up a new ground of challenge. (tl;dr – it basically works like this) foxglove.org.uk/news/grading-a…
First, the position was that the algorithm was sacrosanct and more reliable than mock exams or teachers’ assessed grades.
Ofqual then changed tack, saying mock exams could be more reliable than the algorithm.
THREAD: hundreds of you emailed in anguish about the unfair grading algorithm. This case affects you all, and we are asking for crowdfunding support (gofundme.com/f/fairgrades20…), so we wanted to post our letter as soon as possible. It’s at foxglove.org.uk/news/grading-a…. Key points:
Ofqual just hasn’t got the power to mark students based on an algorithm which actively ignores individual student performance for cohorts over 15. In law-speak, this is called ‘ultra vires’.
The algorithm is irrational – it treats teachers’ rankings as sacrosanct, yet teachers’ assessed grades are treated as either everything (under 5) or nothing (over 15).
Thousands of A-level students get their “results” tomorrow. Covid-19 meant exams were cancelled. So, instead, their grades will in many cases have been generated by a hastily-built government algorithm.
We’ve got concerns about this algorithm.
Firstly, as is so often the case, there’s a lack of transparency about what this algorithm does and how it works.
This is unacceptable. Students have a right to understand how decisions which affect them are made. And algorithmic decision-making needs to be open to scrutiny.
Secondly, it appears the algorithm grades schools rather than students: “Where a subject has more than 15 entries in a school, teachers’ predicted grades will not be used”.
So an individual student’s life chances hang on an estimate based on their school’s historic performance.