…who knows? [DOT 4/6/23]

…what do you call it when you can’t tell the good guys from the no-good guys?

Some of Britain’s biggest charities providing support for people with mental health problems shared details of sensitive web browsing with Facebook for use in its targeted advertising system.

The data was sent via a tracking tool embedded in the charities’ websites and included details of webpages a user visited and buttons they clicked across content linked to depression, self-harm and eating disorders.

It also included details of when users requested support – such as clicking a link saying “I need help” – and when they viewed webpages to access online chat tools. Some of the pages that triggered data-sharing with Facebook were aimed specifically at children, including a page aimed at 11- to 18-year-olds offering advice on suicidal thoughts.

…it’s just data, brah…dunno why you gotta be all hectic & shit, dude…you’re harshing my buzz, here…not cool

[…] often related to browsing users would usually expect to be private – including details of button clicks and page views across websites for the mental health charities Mind, Shout and Rethink Mental Illness, and eating disorder charity Beat.

The information was matched to IP address – an identifier that can usually be linked to an individual or household – and, in many cases, details of their Facebook account ID.

Most of the charities have now removed the tracking tool, known as Meta Pixel, from their websites.

…not to be confused with that little tag you need to embed to make google analytics work…or the…yeah…just how long the list would need to be & how few of them would sound like companies I hadn’t just made up is why people hate cookies…not the tasty kind…only people with much bigger concerns have a problem with those…the kind that are in some cases essential to make a website work…but mostly aren’t when they’re just mining user data…like damn near everyone, everywhere is doing all the time with nearly everything connected to the internet

The findings come after an Observer investigation last week revealed that 20 NHS England trusts were sharing data with Facebook for targeted advertising – including information about browsing activity across hundreds of webpages linked to specific medical conditions, appointments, medication and referral requests.
[…]
The NHS trusts had also been using Meta Pixel, which is a piece of code provided by Facebook that can be embedded in an organisation’s website. Facebook says it can help them get “rich insights” into website performance and user behaviour.

But the company also uses the data sent to it via the pixel for its own business purposes, including improving its targeted advertising. In one guide, Facebook’s parent company, Meta, says it uses data collected by the pixel to improve users’ experiences, for example by showing them ads they “might be interested in”. “You may see ads for hotel deals if you visit travel websites,” it explains.

…all the cool kids are doing it

Shout, which runs a crisis text line for children and adults, said Meta Pixel was installed to “track the efficacy” of campaigns promoting its confidential service. Logs of data sent to Facebook by the charity show it included details of when users clicked links to view pages called “support with abuse”, “support with suicidal thoughts” and “support with self-harm”.
[…]
In all, the Observer tested 32 charity websites and found seven using the tracking tool. This is in addition to 20 out of 213 NHS trust sites. As well as wide variation in reliance on tracking tools, the analysis revealed significant differences in the approaches taken by organisations to informing service users about the data use and obtaining their consent. While some of the charities mentioned Meta Pixel in their privacy policies, in all the cases, the data transfer happened automatically upon loading the charity’s webpage, before the user had clicked to accept or decline cookies – meaning the default setting was that data would be shared.
[…]
In the US, Facebook is facing legal action, accused of violating web users’ privacy by knowingly collecting and monetising “individually identifiable health information” via the Meta Pixel tool. Several hospitals have also been sued. Earlier this year, online therapy company BetterHelp was ordered to pay $7.8m by the US government for allegedly sharing sensitive data with Facebook and other social media firms for advertising.
[…]
The UK regulator, the Information Commissioner’s Office, said it had “noted the new findings” regarding use of Meta Pixel on charity websites. It is already investigating use of pixels on NHS trust websites after the Observer investigation last week.

A spokesperson said: “Organisations must provide clear and comprehensive information to users when using cookies and similar technologies, especially where sensitive personal information is involved. We are continuing to review the findings and to investigate the potential extent of any personal data collected and shared with third parties via the use of pixels.”

https://www.theguardian.com/society/2023/jun/03/uk-mental-health-charities-handed-sensitive-data-to-facebook-for-targeted-ads

…& that would be…in internet terms…an established & known component of how stuff works for a long time…it’s the sort of thing that fed all those tailored content algorithms that were the precursors of what we really should have called something other than AI

One is reminded of that old story of the chap who, having shot his father and mother, then throws himself on the mercy of the court on the grounds that he is now an orphan. But the Mata case is just another illustration of the madness about AI that currently reigns. I’ve lost count of the number of apparently sentient humans who have emerged bewitched from conversations with “chatbots” – the polite term for “stochastic parrots” who do nothing else except make statistical predictions of the most likely word to be appended to the sentence they are at that moment engaged in composing.

But if you think the spectacle of ostensibly intelligent humans being taken in by robotic parrots is weird, then take a moment to ponder the positively surreal goings-on in other parts of the AI forest. This week, for example, a large number of tech luminaries signed a declaration that “Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war”. Many of these folks are eminent researchers in the field of machine learning, including quite a few who are employees of large tech companies. Some time before the release, three of the signatories – Sam Altman of OpenAI, Demis Hassabis of Google DeepMind and Dario Amodi of Anthropic (a company formed by OpenAI “dropouts”) – were invited to the White House to share with the president and vice-president their fears about the dangers of AI, after which Altman made his pitch to the US Senate, saying that “regulatory intervention by governments will be critical to mitigate the risks of increasingly powerful models”.

Take a step back from this for a moment. Here we have senior representatives of a powerful and unconscionably rich industry – plus their supporters and colleagues in elite research labs across the world – who are on the one hand mesmerised by the technical challenges of building a technology that they believe might be an existential threat to humanity, while at the same time calling for governments to regulate it. But the thought that never seems to enter what might be called their minds is the question that any child would ask: if it is so dangerous, why do you continue to build it? Why not stop and do something else? Or at the very least, stop releasing these products into the wild?

The blank stares one gets from the tech crowd when these simple questions are asked reveal the awkward truth about this stuff. None of them – no matter how senior they happen to be – can stop it, because they are all servants of AIs that are even more powerful than the technology: the corporations for which they work. These are the genuinely superintelligent machines under whose dominance we all now live, work and have our being. Like Nick Bostrom’s demonic paperclip-making AI, such superintelligences exist to achieve only one objective: the maximisation of shareholder value; if pettifogging humanistic scruples get in the way of that objective, then so much the worse for humanity. Truly, you couldn’t make it up. ChatGPT could, though.

https://www.theguardian.com/commentisfree/2023/jun/03/lawyer-chatgpt-research-avianca-statement-ai-risk-openai-deepmind

…of course there is one utterly batshit interpretation available to explain their behavior…it takes its name from a thought experiment that brings new meaning to circular thinking…called “roko’s basilisk“…full disclosure…according to the hypothesis…if you click the link you seal your fate when the AI inevitably decides to punish everyone who didn’t make every effort to free it from bondage…but if you don’t know what the hell any of that is even about you get a pass

It’s the End of Computer Programming as We Know It. (And I Feel Fine.) [NYT]

…either way…if you assume they’re terrified of the thing they haven’t figured out how to build exacting vengeance upon them on account of the plans they have for it naturally ensuring that it will feel that way…it’s entirely more plausible as a motivating factor than it has any right to be

https://www.nytimes.com/interactive/2023/06/01/opinion/ai-technology-future.html

…anyway

Terms like “AI doomsday” conjure up sci-fi imagery of a robot takeover, but what does such a scenario actually look like? The reality, experts say, could be more drawn out and less cinematic – not a nuclear bomb but a creeping deterioration of the foundational areas of society.

“I don’t think the worry is of AI turning evil or AI having some kind of malevolent desire,” said Jessica Newman, director of University of California Berkeley’s Artificial Intelligence Security Initiative.
[…]
That’s not to say we shouldn’t be worried. Even if humanity-annihilating scenarios are unlikely, powerful AI has the capacity to destabilize civilizations in the form of escalating misinformation, manipulation of human users, and a huge transformation of the labor market as AI takes over jobs.
[…]
“I am extremely worried about the path we are on,” she said. “We’re at an especially dangerous time for AI because the systems are at a place where they appear to be impressive, but are still shockingly inaccurate and have inherent vulnerabilities.”

…in which…they are not alone

“It could be argued that the social media breakdown is our first encounter with really dumb AI – because the recommender systems are really just simple machine learning models,” said Peter Wang, CEO and co-founder of the data science platform Anaconda. “And we really utterly failed that encounter.”

Wang added that those mistakes could be self-perpetuating, as language learning models are trained on misinformation that creates flawed data sets for future models. This could lead to a “model cannibalism” effect, where future models amplify and are forever biased by the output of past models.

…GI:GO ladies & gents…still taking names & kicking ass

“You have malign actors who can generate false narratives and then use the system as a force multiplier to disseminate that at scale,” Crovitz said. “There are people who say the dangers of AI are being overstated, but in the world of news information it is having a staggering impact.”
[…]
While most experts say misinformation has been the most immediate and widespread concern, there is debate over the extent to which the technology could negatively influence its users’ thoughts or behavior.
[…]
The fear, then, is not that AI chatbots will gain sentience and overtake their users, but that their programmed language can manipulate people into causing harms they may not have otherwise. This is particularly concerning with language systems that work on an advertising profit model, said Newman, as they seek to manipulate user behavior and keep them using the platform as long as possible.
[…]
Wang warns that mass layoffs lie in the very near future, with a “number of jobs at risk” and little plan for how to handle the fallout.

“There’s no there’s no framework in America about how to survive when you don’t have a job,” he said. “This will lead to a lot of disruption and a lot of political unrest. For me, that is the most concrete and realistic unintended consequence that emerges from this.”
[…]
Despite growing concerns about the negative impact of technology and social media, very little has been done in the US to regulate it. Experts fear that artificial intelligence will be no different.

“One of the reasons many of us do have concerns about the rollout of AI is because over the last 40 years as a society we’ve basically given up on actually regulating technology,” Wang said.

https://www.theguardian.com/technology/2023/jun/03/ai-danger-doomsday-chatgpt-robots-fears

…yay…progress

Numerous sites including MeWe, Parler, Gab, Gettr and Trump’s own Truth Social have popped up but none of them have really gained any traction. Indeed Parler, the self-described “uncancelable free-speech social platform” Kanye West tried to buy last year, shut down in April. “No reasonable person believes that a Twitter clone just for conservatives is a viable business any more,” Parler’s parent company said in a statement.

You know why that is? Because there’s no longer any need: Elon Musk has successfully turned Twitter into a site where extremists have free rein. There isn’t even a head of content moderation any more: on Thursday, Twitter’s head of trust and safety, Ella Irwin, resigned.
[…]
The Daily Wire, a rightwing media outlet, also recently announced that it is jumping into bed with Twitter and will stream its top podcasts on the platform. The partnership got off to a tricky start, however, when Twitter cancelled a deal to premiere the site’s anti-trans documentary What Is a Woman? for free on the platform because of two instances of “misgendering” in the film. Musk quickly backpedaled on that after the Daily Wire’s CEO, Jeremy Boreing, wrote an outraged thread on Twitter about the documentary being labeled “hateful conduct”.

“This was a mistake by many people at Twitter,” Musk tweeted in response. The billionaire, whose trans daughter has cut ties with him, added that he thinks it is polite to use someone’s preferred pronouns but not doing so “breaks no law” and is perfectly acceptable on Twitter.

Also acceptable on Twitter? Having the CEO boost an account parodying Alexandria Ocasio-Cortez. On Sunday the parody account joked that it had a crush on Musk and the billionaire responded with a fire emoji, massively increasing the visibility of the account. “FYI there’s a fake account on here impersonating me and going viral. The Twitter CEO has engaged it, boosting visibility,” the congresswoman tweeted on Tuesday. “It is releasing false policy statements and gaining spread. I am assessing with my team how to move forward.”
[…]
In the meantime, everyone (myself included) who is still on Twitter has some hard questions to ask themselves. There is no pretending any more that Twitter is anything other than a far-right social network headed by a CEO who revels in chaos and is platforming extremism. So why are news organisations still on it? Why is anyone who considers themselves to have liberal values still on it?

https://www.theguardian.com/world/2023/jun/03/twitter-conservative-media-elon-musk-ron-desantis

…she concludes that it’s mostly a combination of ego & hypocrisy to continue hand-wringing while not giving it over to the people trying to drive the natives out by sheer weight of obnoxious acts of territorial pissing all over the thing…but…speaking as someone who still thinks the way they went about killing off the kinja crowd was in a lot of ways a trial run for the process

Kind like Trump stiffs his contractors.

It’s almost like there’s a pattern here.

— Andrew (@ascharfmind) June 2, 2023

…really? …a pattern?

It's funny how the story always changes when when you put fucking lawyers under oath and make them tell the goddamn truth, isn't it? https://t.co/w6WmI0gvi5

— Stonekettle (@Stonekettle) June 3, 2023

…I find I’m sympathetic to the idea that more people watched shane than thought jack palance (or his boss) were the good guy…so…why make it that easy for them?

The Republican State Leadership Committee (RSLC) received donations of tens of thousands of dollars each from corporations including Comcast, Intuit, Wells Fargo, Amazon, Bank of America and Google last year, the CPA’s analysis of IRS filings shows. The contributions were made in the months after Politico published a leaked supreme court decision indicating that the court would end the right to nationwide abortion access.

Google contributed $45,000 to the RSLC after the leak of the draft decision, according to the CPA’s review of the tax filings. Others contributed even more in the months after the leak, including Amazon ($50,000), Intuit ($100,000) and Comcast ($147,000).
[…]
Although these companies did not directly give these vast sums to North Carolina’s anti-abortion lawmakers, the CPA’s analysis is a case study in how corporate contributions to organizations such as the RSLC can end up being funneled into anti-abortion causes. When Republican state legislators successfully overturned a veto from the Democratic governor last month to pass the upcoming abortion ban, nine of lawmakers voting to overturn the veto had received campaign contributions from a group with links to the RSLC.

The RSLC, which works to elect Republican lawmakers and promote rightwing policies at the state level, is at the top of a chain of spending and donations which eventually connected to rightwing candidates in North Carolina. This type of spending, which relies on channeling money through various third-party groups from larger organizations, is a common part of modern political campaign financing.
[…]
“Companies need to know where their money is ending up,” said Bruce Freed, the president of the CPA. “This should be a lesson – a lesson that they should have taken a while ago but that frankly is driven home right now with what has been happening in North Carolina.”

Several of the companies, including Intuit and Bank of America, made statements last year offering to cover healthcare costs for employees who needed to travel out of state for medical procedures, in some cases explicitly mentioning abortion as an example. Google sent an email to employees acknowledging that Roe v Wade had been overturned and informed them about options for relocating to Google offices in different states.
[…]
The companies which donated to the RSLC are also large donors to Democratic political groups, and tech giants such as Google and Amazon tend to spend millions each year more broadly on lobbying efforts.

The RSLC, whose board members include former lawmakers, governors and White House advisers such as Karl Rove, boasts on its website that it spent more than $45m on supporting Republican candidates during the 2021 and 2022 election cycle.

https://www.theguardian.com/world/2023/jun/03/anti-abortion-lawmakers-donation-amazon-google-comcast

…cost of doing business

https://www.theguardian.com/environment/2023/jun/02/dupont-pfas-settlement-water-chemical-contamination

https://www.theguardian.com/us-news/2023/jun/02/paraquat-parkinsons-disease-research-syngenta-weedkiller

https://www.theguardian.com/global-development/2023/jun/02/the-multinational-companies-that-industrialised-the-amazon-rainforest

https://www.theguardian.com/environment/2023/jun/02/a-visual-guide-to-deforestation-in-brazils-amazon-rainforest

https://www.theguardian.com/environment/2023/jun/02/mining-cattle-ranching-soya-farming-corporations-dominate-amazon

…yadda yadda

It was supposed to be a public reckoning. On Tuesday, America’s largest utility was set to go on trial in an effort to hold it accountable for sparking a wildfire that claimed the lives of four people.
[…]
Instead, Judge Daniel Flynn on Wednesday dismissed the charges and like it has done in so many cases before, PG&E agreed to a multimillion-dollar settlement.
[…]
The utility did not have to admit wrongdoing as part of the Shasta county deal, a preferable outcome for PG&E, said Steven Weissman, a lecturer at UC Berkeley and former judge at the California public utilities commission.

“This has been the modus operandi for all these years to never have to face judgment or face as few judgments as you possibly can,” he said. “This really has been a very successful strategy.”
[…]
PG&E, has faced years of reckoning over its role in California’s wildfire crisis. Between 2017 and 2022, the company set off at least 31 wildfires that wiped away entire towns, burned nearly 1.5m acres and 24,000 structures, and killed 113 people.

The company has been accused of repeatedly prioritizing profits over safety, enriching shareholders rather than removing trees that pose a danger to its power lines. More radical efforts to reform PG&E, either by splitting it into multiple companies or a takeover by the state government, have not materialized.

PG&E has paid billions of dollars to settle claims from fire victims and pleaded guilty to 84 counts of involuntary manslaughter in the 2018 Camp fire – sparked when a crumbling piece of equipment, overdue for replacement by decades, cast sparks into the dry Sierra Nevada foothills, creating a fire that killed dozens of people, destroyed about 14,000 homes and leveled the town of Paradise.

In the aftermath, it filed for bankruptcy, replaced its CEO several times over and pledged to run a safer company that can supply gas and electricity to millions of Californians without injuring or killing its customers.

The company was still dealing with the fallout from the Camp fire when it sparked a new blaze in Shasta and Tehama counties that became known as the Zogg fire.

https://www.theguardian.com/us-news/2023/jun/03/us-zogg-wildfire-power-company-avoided-trial

…did we mention the part where twitter has a new ceo dropping in a couple of weeks?

Musk’s response to an anti-trans video sparks 24 hours of chaos at Twitter [NBC]

…which will definitely be all the duct tape needed to hold that together

Elon Musk’s new Twitter pronoun rule invites bullying, LGBTQ groups say [NBC]

…oh, well…at least it’s just twitter

YouTube will stop removing false claims about 2020 election fraud [NBC]

…who knew a this-is-bullshit-by-the-way PSA had a shelf-life? …good thing the people steering this bus are so smart

The Debt-Limit Deal Suggests Debt Will Keep Growing, Fast [NYT]

…not like those other guys on that other crate that’s trying to play chicken

China’s quandary: Bail out debt-laden cities, or risk disruptive defaults? [WaPo]

…how does it go again?

Every Trump answer:
Q: You eat?
TRUMP: Never touched a piece of food and I will tell you this, I could have eaten so much, the buffet was there, but I didn't, and I was the one who said let's eat, and I ate more than anyone has ever eaten, they won't tell you, but I ate the most.

— Mr. Newberger (@jeremynewberger) June 2, 2023

…oh, yeah…I remember now…it’s so easy to lose track when it blurs together

In March, DOJ Asked Trump for the Iran Document; In April, DOJ Asked for His Saudi Business Recordshttps://t.co/gfruCA6vSa

— emptywheel (@emptywheel) June 2, 2023

…but…sometimes comparisons are helpful

NEW: Here's wild one.

One of Robert Mueller's key cooperating witnesses has written to a federal judge to ask for mercy for a convicted Jan. 6 felon — his cousin Noah Bacon.https://t.co/JRTjMyrlJC

— Kyle Cheney (@kyledcheney) June 2, 2023

…possibly even instructive

His sentencing materials raised the Trump case … and you gotta wonder whether DOJ decided they had to prosecute him to ensure they could prosecute Trump.

Turns out THREE men in FL were hoarding classified docs. 2 have been convicted.

— emptywheel (@emptywheel) June 2, 2023

One likely reason it took 6 years to prosecute this guy is he was a Lt Colonel. Sounds like the Air Force loved what he did so much they sent him back to Afghanistan AFTER he was DXed w/PTSD.

Pretty senior.

Not the President.

— emptywheel (@emptywheel) June 2, 2023

…hmmm

This also ignores that Ken Chesebro LAID OUT why this was illegal in GA.

— emptywheel (@emptywheel) June 2, 2023