OfDIA/DSIT. Ex- Privacy International. African Studies PhD; writes on #digitalID. GoodID community champion. Views my own. @drsamaki.bsky.social

London, UK
Joined November 2014
It was my 10th anniversary of joining Twitter last week; it’s the only social media platform I’ve ever really engaged with. But I guess when it’s time to go, it’s time to go. You can find me @drsamaki.bsky.social
1
101
While the trolley problem took off in the public imagination, it seems that there are many thought experiments in the Philosophy of Mind and Philosophy of Language that seem relevant to our current situation; with a lot of ants doing a lot of caricatures of Winston Churchill. 1
1
2
166
But: what does it then mean for the status of a conversation with a chatbot? It doesn’t seem to reach any kind of standard of reference to the world (real or imaginary). So is there nothing going on here except the user anthropomorphising the chatbot? 3
1
70
I need a nice, easy guide to the Philosophy of recent developments in AI that’ll stop me worrying about all of this, as I know enough for it to bother me but not enough to reach any conclusions. Any pointers? 4
1
1
114
I'm not convinced a ban will do much to stop young people from accessing social media; but it will make it far harder to make sure they do so safely.
It’s an unworkable idea and one that restricts young people’s freedoms, inc access to information and resources and yes, even friendships. Resist.
2
1
1
367
Wouldn't worry too much if you'll be in London that day, as they're heading out of the city.
Are they planning to rearrange the entire landscape of the city?
72
The injustice of requiring identification to attend school: it takes a societal harm falling upon the broad shoulders of the state(‘ghost children’ and corruption in schools) and makes it an individual harm (kids from marginalised backgrounds can’t attend school).
1
1
2
152
When the starting point of your DPI is identity and identification: you’re running the risk of these types of harms at every stage of the system.
33
Sweden and Norway are realising that their cashless systems are vulnerable to Russian cyberattacks, and so are back to encouraging the use of cash. theguardian.com/world/2024/o… In other news: the push for DPI around the world continues regardless.
1
13
2
20
1,579
I don't know if @Suitpossum is still around Twitter, but it's interesting to note that Sweden and Norway are starting to do the right thing (not least, for the hundreds of thousands excluded from these systems); all it took was an unprecedented global crisis to do it.
3
69
This is unbelievably lazy and dangerous, especially from a former solicitor running to lead the Conservative Party Robert Jenrick gets the charges wrong then launches into "cover-up" dogwhistle while surely knowing how the law limits what information can be released before trial
The public had a right to know the truth straight away. I am seriously concerned that facts may have been withheld from the public. Keir Starmer must urgently explain what he knew about the Southport attack and when he learned it.
218
495
47
2,093
552,848
Replying to @TomSamaki
The DVS section is unsafe as it has too many ministerial powers. I suggest amending DVS so an independent party (@ICOnews??) develops a statutory Code of Practice to incorporate the officially approved identity principles into the DVS section. See gov.uk/government/publicatio…
2
2
462
The analysis we were all waiting for; and - gosh, there's some nasty things lurking in that bill. The right not to be subjected to automated decision making; and changes to the use of data for the Research, Archive and Statistical purposes leap out at me.
What are the main features of the Data (Use and Access) Bill? The blog is in three parts: what has been dropped from DPDI in DUAB; what has been copied from DPDI into DUAB; and what is new to DUAB. Details at end of the blog. amberhawk.typepad.com/amberh…
78
The other advantage of existing automated transcription is that the errors are usually very obvious for anyone reading the transcript, as opposed to OpenAI's hallucinations.
The extremely dumb thing about this, like your spellchecker getting worse, is that we had perfectly good specialized models for this purpose that were quite effective and now everyone is paying more for non specialized models for the same tasks, that are worse.
2
4
807
My knowledge of Habemas' Discourse Ethics is an undergraduate course twenty years ago; and while @JGFinlayson did his best, it wasn't my finest hour. But i can still say this AI "Habemas Machine" is a problematic use of the name. arstechnica.com/ai/2024/10/g…
101
Looks to me that the Government is trying to use "tough on crime" language, via The Sun... But also playing us for fools, and hoping we don't realise that facial recognition cameras don't only record criminals.
They’re turning Britain into an open prison. “Virtual jails”: released prisoners could be tracked with facial recognition Insider said “we need to learn from surveillance states who spy on their own citizens & use the same tactics with our criminals”😱 thesun.co.uk/news/31245640/f…
1
1
86
I love this about Musk and politics: he owns a social media network (probably one of the most powerful tools for shaping public opinion in history); but makes such a hash of it that he has to resort to literally giving people money.
69
Tom Fisher retweeted
If you don’t feel stupid doing research you’re not researching hard enough
39
2,606
269
14,045
372,507