gAlienLifeform

JUHASZ: DiBenedetto now works for Louisiana's Department of Education and is in charge of bringing Amira into more classrooms. He says by the time the state's two-year pilot is over...

DIBENEDETTO: I think we're going to see some interesting impacts, and we'll definitely have some data to make prudent decisions in the future.

JUHASZ: Like whether to spend even bigger money on AI. The company behind Amira says 2 million children already use the tool. Experts caution the technology isn't a replacement for teachers or even all tutors. It can't build relationships with students like humans can.

MONTAGNINO: I'm old-school. I still believe people, especially with reading for little kids - that's where it's at.

JUHASZ: Montagnino, the principal in Gretna, says for that reason, she was skeptical at first.

MONTAGNINO: But this, to supplement good science of reading instruction in the classroom? This is great.

JUHASZ: And it's likely to get better because just as kids are learning from Amira, it's learning from them, too.

[Bolding added]

So it seems an alternative headline for this story would be "Private for profit company gets paid to collect training data for its AI from children who could face disciplinary or legal consequences for non-compliance"

Archived at https://archive.is/Th1Sq

In June, the Texas Department of Public Safety (DPS) signed an acquisition plan for a 5-year, nearly $5.3 million contract for a controversial surveillance tool called Tangles from tech firm PenLink, according to records obtained by the Texas Observer through a public information request. The deal is nearly twice as large as the company’s $2.7 million two-year contract with the federal Immigration and Customs Enforcement (ICE).

Tangles is an artificial intelligence-powered web platform that scrapes information from the open, deep, and dark web. Tangles’ premier add-on feature, WebLoc, is controversial among digital privacy advocates. Any client who purchases access to WebLoc can track different mobile devices’ movements in a specific, virtual area selected by the user, through a capability called “geofencing.” Users of software like Tangles can do this without a search warrant or subpoena. (In a high-profile ruling, the Fifth Circuit recently held that police cannot compel companies like Google to hand over data obtained through geofencing.) Device-tracking services rely on location pings and other personal data pulled from smartphones, usually via in-app advertisers. Surveillance tech companies then buy this information from data brokers and sell access to it as part of their products.

WebLoc can even be used to access a device’s mobile ad ID, a string of numbers and letters that acts as a unique identifier for mobile devices in the ad marketing ecosystem, according to a US Office of Naval Intelligence procurement notice.

Wolfie Christl, a public interest researcher and digital rights activist based in Vienna, Austria, argues that data collected for a specific purpose, such as navigation or dating apps, should not be used by different parties for unrelated reasons. “It’s a disaster,” Christl told the Observer. “It’s the largest possible imaginable decontextualization of data. … This cannot be how our future digital society looks like.”

Archived at https://web.archive.org/web/20240827115133/https://www.texasobserver.org/texas-dps-surveillance-tangle-cobwebs/

Archived at https://web.archive.org/web/20240826162608/https://www.technologyreview.com/2024/08/14/1096534/homeland-security-facial-recognition-immigration-border/

Inconsistencies in phone call records; a confusing time stamp on a Google search to learn how long it would take for a person to die in the cold; health data that showed a person descending a stairway — or maybe in a car.

While some forensic work is well established, such as DNA evidence, other technologies aren’t quite as grounded, as the Read trial showed. In particular, the field of digital forensics continues to evolve, shaped by court challenges and advancing technology. So, questions around the validity of that data have become the latest frontier in what legal observers call the “battle of experts”: dueling interpretations of an unsettled science.

And, with enough legal prowess — and financial resources — defendants can line up parades of experts to try to undermine a prosecution witness’s interpretation of forensic data, from the timing of a Google search to the movement of a human body.

Archived at https://web.archive.org/web/20240814121648/https://www.bostonglobe.com/2024/08/12/metro/karen-read-digital-forensics/

Archived at https://web.archive.org/web/20240722215141/https://www.dailydot.com/debug/kamala-twitter-account-follow-limit-elon-musk/

Archived at https://web.archive.org/web/20240603122116/https://www.npr.org/2024/06/02/nx-s1-4985877/25-years-ago-napster-changed-how-we-listen-to-music-forever

Zhang, an electrical engineer in Boston, decided to post about trying to unlock his Justice Tech Solutions Securebook 5 on the social platform X. The thread went viral — also catching the attention of Washington corrections officials, who have used the device for college programming since 2020.

Of particular concern was an article about Zhang’s thread published on a hacker website that shared the default password for the underlying software that starts the laptop’s operating system, presenting what the Department of Corrections considered a security concern.

The department then announced Thursday, five days after Zhang’s viral post, that it would collect all secure laptops from incarcerated students statewide “to provide an immediate system update.” By Saturday, corrections staff had collected around 1,200 laptops, spokesperson Chris Wright said in an email.

Wright confirmed no one incarcerated in Washington prisons had attempted to unlock their devices but said the decision was “made out of an abundance of caution.” It wasn’t immediately clear whether other states whose corrections departments use Securebook 5 laptops have also pulled the devices.

Archived at https://ghostarchive.org/archive/LS3co

e; updated the title due to popular demand

Archived at https://web.archive.org/web/20240215123711/https://www.brennancenter.org/our-work/research-reports/closing-data-broker-loophole

I haven't thought enough about it to endorse these ideas, but it seems like a really interesting discussion and one the open source development community ought to be thinking about

A few ideas for what this could look like:

• A modern, content-focused subset of HTML and CSS. I think companies should be trusted to brand and promote themselves, and the failure to make this part of the RSS specification may be one factor that might have turned off publishers from investing more deeply into RSS. So, let’s give it to them in the form of some basic design, including access to fonts, graphics, and simple layout intended for a narrow space. The content should be static, to be clear—no JavaScript here—but it should allow for enough flexibility that if people want to experiment, they can. We already have an existing spec that does much of this—the open-source AMP standard, developed primarily by Google—though I understand if we don’t want to use it, due to its controversial history. Whatever this theoretical looks like, it should be flexible and easy for end users to implement.
• Features to encourage use of rich text. Adding features like data visualizations, graphics, and embedded video that are not part of the regular RSS specification could add appeal to this new format by offering something that newsletters do not have, while giving it advantages over a standard RSS feed.
• Built-in access management. If you, as a publisher, want to gate all or part of a feed item, you can do so, and offer your own integration as to how to resolve the block. Essentially, build subscriptions or regwalls directly into the feed—and make it so that you don’t have to work with a middleman like a Substack to do so. Don’t want the bots or the LLMs to access your life’s work? Build in a regwall.
• Built-in integrations for distribution. RSS is built for distribution, but I wonder if this new thing I’m suggesting should talk ActivityPub, or easy to distribute in a newsletter format for people who still want to read in their inbox. Make it so that people can follow you wherever they’re comfortable, rather than being forced to read in a newsletter format, or a social media format.
• Limited, but useful, analytics. You should know how many people read your newsletters, and you should know how they’re read, but you probably shouldn’t know much else. Podcasting has benefited from a lack of data poisoning the well—and honestly, resetting the conversation around data could really help strengthen the content ecosystem at this juncture.

Archived at https://web.archive.org/web/20240207155759/https://tedium.co/2024/02/06/rss-creator-economy-rethink/

Archived at https://web.archive.org/web/20240206132148/https://www.vice.com/en/article/m7bwp4/axon-acquires-fusus-ai-surveillance-retail-healthcare

Archived at https://archive.is/nB7Db