Reuben Binns @[email protected]

computery guy @CompSciOxford, @KelloggOx. HCI; ML; privacy; security; political econ of tech. [email protected]

#THFC

he/him
[email protected]

[Reuben Binns @[email protected] @RDBinns]

I'm going to be tweetin' less here and tootin' more over on the Fediverse - find me at [email protected] for EXCLUSIVE CONTENT

[Lilian Edwards @lilianedwards]

New paper by @rdbinns and myself on why genAI hallucinates and its implications in terms of libel and data protection is now strangely relevant after David Mayer-gate 🤣 where individuals may or may have used law to remove themselves fromChatGPT https://t.co/oHizdMopT7

[Alex Chalmers @chalmermagne]

new from me: I’ve teamed up with @anastasiabekt to explore why so much policy work is basically pointless. We walk through the mistakes tech companies, consultancies, think tanks, and lobbyists make, how it impoverishes debate, and how we can improve it. 🧵

[Lina Khan @linamkhan]

1. Firms that lure workers with false earnings claims are breaking the law.

@FTC has taken action against @Lyft for deceiving drivers about how much they could expect to earn on its platform.

We've ordered Lyft to stop this conduct & pay $2.1 million.

https://t.co/oQF1KTk3WO

[Aaron Benanav @abenanav]

If you want to see the other graph you need, to understand our world, it is this one. It depicts the rising wealth/income ratio. Basically, the value of all wealth (businesses, real estate, financial assets, etc) is rising much faster than all income (which is just GDP).

[konrad kollnig @FascinatingTech]

🚨PhD student + team player sought. Please share, repost, or apply.. Are you passionate about how digital technologies affect citizens, curious about the social element of those technologies, and eager to work in a truly interdisciplinary team? Then, you may want to consider joining us in the #RegTech4AI project at the Law & Tech Lab in the beautiful city of Maastricht (Netherlands). As of this month, we already have a team of four outstanding interdisciplinary researchers, trying to define the future of AI & Digital Technologies.

Link to application site: https://t.co/wk6KiwWQGn

Project details: https://t.co/cN8gWoqkwb

[Reuben Binns @[email protected] @RDBinns]

You'll come a-texting Matilda with me

[10 News Sydney @10NewsSyd]

NSW Ambulance has completed a dramatic rescue, retrieving a woman who was stuck upside down between rocks in the Hunter Valley.

Matilda Campbell dropped her phone between the rocks while walking with friends, and when she attempted to retrieve it she fell and became trapped, hanging by her feet.

She was there for seven hours, and specialist rescue paramedics had to remove boulders weighing up to 500 kilograms to free her.

Campbell was eventually freed and only sustained minor scratches and bruises, but unfortunately her phone did not survive.

[Owen Boswarva @owenboswarva]

I doubt any of this is true. At most these comments highlight a woeful lack of expertise in Whitehall on application of GDPR/DPA and the common law duty of confidentiality. All ICO really did was provide assurances that data protection wasn't a barrier.
https://t.co/zoMwHhOZK2

[Worker Info Exchange @WorkerInfoX]

From our Dying For Data webinar, @RDBinns explains how dynamic pricing is affecting drivers’ pay.

Watch the full video on Youtube https://t.co/gLpcxRA2tc

[@frederike.bsky.social @F_Kaltheuner]

NEW: Should Europe invest billions in A𝗜?
Without redirecting the current trajectory, Europe's industrial policy risks further entrenching Big tech's dominance. But there are alternative pathways forward!

This collection of essays by @AINowInstitute sets out to provide EU policy makers with policy research, perspectives, and evidence about the challenges of expanding public investment in a highly concentrated global AI market.

https://t.co/dmpEpFSFI7

Curated by: @F_Kaltheuner @SaariLeevi @ambaonadventure @sarahbmyers

Contributors incl: @francesca_bria @CeciliaRikap
@sarahchander @ZWarso @Caffar3Cristina @sedyst
@margaridasilva2 @JeroenMerk @kimvsparrentak @burcuno @thornet @udbhav_tiwari @franzbonf @FiekeJ @markscott82

[The Marshall McLuhan LP @fireh9lly]

Awesome, it screws over everything that the UK is good at (arts) in favour of some tech billionaires from the US who don't even pay taxes here

[Elton Reichert @Eltoninfosec]

Conflict of interest (COI) leads to a strong downward valuation and credibility of research output

@zingales @TomValletti @LancieriFilippo @JoshuaYLevy @Barrios_Econ Shashank Singh

The paper advocates for a clearer COI policy: the profession’s #me-consult-too moment

[Reuben Binns @[email protected] @RDBinns]

@Angronn1 lucky you!

[Reuben Binns @[email protected] @RDBinns]

Maybe people who believe LLMs can reason are actually using a different meaning of 'reason', which is exactly the same as ours except it also includes pattern matching by LLMs 🤔🤔🤔

[Reuben Binns @[email protected] @RDBinns]

'LLMs can't really reason, they just pattern match'

Wittgenstein: 'ah, but how do we know they aren't following their own private reasoning rules which are the same as ours, except when the reasoning problem has superficial differences to their training data?'

[Reuben Binns @[email protected] @RDBinns]

'LLMs can't really reason, they just pattern match'

Wittgenstein: 'ah, but how do we know they aren't following their own private reasoning rules which are the same as ours, except when the reasoning problem has superficial differences to their training data?'

[electronic Max @emax]

"A new paper from Apple draws attention to a persistent problem in LLMs: their reliance on pattern matching rather than genuine logical reasoning. As a result, adding irrelevant information—can lead to vastly different answers from the models."
https://t.co/OB5V7uMIep

[Leon Yin @LeonYin]

@obiriders NYC has a min wage formula that penalizes companies for inefficiency. The busier drivers seem, the less each ride is worth. To keep wages low, Uber and Lyft inflate driver efficiency stats by erasing idle time. We estimate millions at stake monthly.
https://t.co/W1hI5besZy

[Worker Info Exchange @WorkerInfoX]

📺Watch our webinar replay:
➡️James DiGiovanni on how @nyctaxi use data to protect pay & cut congestion
➡️@RDBinns analyses our data to show harm of Uber dynamic pricing to UK workers
➡️@veenadubal on algorithmic wage discrimination & the political economy
https://t.co/rscThybA2U

[Reuben Binns @[email protected] @RDBinns]

I conclude that political philosophy approaches to algorithmic fairness would benefit from stepping out of the shadow of Rawlsianism, and embracing other paradigms. Available here: https://t.co/UdcWcrUggW and OA preprint here: https://t.co/6wLpYZEe5S

[Reuben Binns @[email protected] @RDBinns]

New short paper: 'If the Difference Principle Won’t Make a Real Difference in Algorithmic Fairness, What Will?'

An invited response to Ulrik Franke's excellent ‘Rawlsian Algorithmic Fairness and a Missing Aggregation Property of the Difference Principle’ https://t.co/E5mRlJKhnY

[Reuben Binns @[email protected] @RDBinns]

These include the situational level (e.g. applying existing non-discrimination law to algorithms), the institutional (e.g. new rules banning high risk systems), and the systemic level (e.g. tech that helps build alternative power structures).