Ian Linkletter

@linkletter.bsky.social

425

Followers

406

Following

116

Posts

Emerging Technology & Open Education Librarian fighting for the right to link.

Posts Replies Media

Today I returned to UBC's Faculty of Education to present "Academic Surveillance in Higher Education" in @surita.bsky.social's "Critical Multicultural Perspectives on Learning with Digital Technologies: Redressing Educational Inequalities Through Open Pedagogy" course. As always, I learned so much.

I created something. It's only the start. www.privacylibrary.ca

Canadian Privacy Library – Open privacy knowledge for allwww.privacylibrary.ca

Today is Day 1400 of Proctorio's lawsuit.

Racial, skin tone, and sex disparities in automated proctoring software. This study shows how facial detection software can be disastrous for students of colour being judged by AI proctoring like #RespondusMonitor and #Proctorio www.frontiersin.org/journals/edu...

Frontiers | Racial, skin tone, and sex disparities in automated proctoring softwarewww.frontiersin.org Students of color, particularly women of color, face substantial barriers in STEM disciplines in higher education due to social isolation and interpersonal, ...

Four years ago today, the CEO of Proctorio posted a UBC student's chat log to Reddit to retaliate against her for posting about shoddy support. I criticized him for this, and after publicly arguing about it with many educators, he apologized and deleted all his tweets. Two months later, I was sued.

Are you feeling moral distress about AI proctoring and detection in your school? Sign the pledge against academic surveillance! “I will do everything in my power to obstruct, work against, and end academic surveillance technologies and their harm against students.” pledge.againstsurveillance.net

#AgainstSurveillance – Pledge Against Academic Surveillancepledge.againstsurveillance.net

University of Colorado Boulder is ending their relationship with Proctorio. 1,403 people signed a petition calling for them to stop the use of Proctorio. Congratulations to all CU Boulder students. 🎉 Petition: www.change.org/p/university... News Link: oit.colorado.edu/services/tea...

Sign the Petitionwww.change.org I am a student and IT student employee at CU Boulder. I and many other students have complained about the privacy, security, and accessibility concerns of using the service Proctorio. This petition is...

Don’t be a Proctorio president. Don’t be a Proctorio provost. Don’t be a Proctorio professor. Stand up for students. Protect them from the harms of AI surveillance in their own homes, which they don’t and can’t consent to. Build a trusting learning environment by being ethical.

Students cannot consent to Proctorio if they do not know how it works. Proctorio files lawsuits to prevent students from knowing how it works. None of Proctorio's competitors sue educators and students, like they have. Protect students: don't be a Proctorio professor. Don't be a Proctorio provost.

Vanderbilt University’s contract with Proctorio has ended. They're going with a surveillance company that doesn't sue educators and students instead.

Reposted by

Ian Linkletter

Put this whole paragraph on repeat.

Writers are losing their jobs after false positives from ai detectors.

AI Detectors Get It Wrong. Writers Are Being Fired Anywaygizmodo.com AI is already stealing writers’ work. Now they’re losing jobs over false accusations of using it.

Reposted by

Ian Linkletter

Writers are losing their jobs after false positives from ai detectors.

AI Detectors Get It Wrong. Writers Are Being Fired Anywaygizmodo.com AI is already stealing writers’ work. Now they’re losing jobs over false accusations of using it.

Is this normal? Proctorio has a new Whistleblower Policy stating that "Proctorio will not intimidate, harass, discharge, threaten, or discriminate or retaliate against" whistleblowers *if* they unilaterally determine "good faith".

Because I tweeted 7 publicly accessible YouTube video links in August 2020, Proctorio has been suing me for the past 1,363 days.

Me: I've heard a lot of bad things about Proctorio. Are they true? ChatGPT: Key points of concern include privacy issues, bias, accessibility, legal controversies, ethical controversies, and technical issues. ChatGPT links to three sources. Link: chatgpt.com/share/ed310a...

chatgpt.com

How strange... Proctorio paid a Substack bro to speak at "ProctorioX" last month and now he's falsely written that I'm the one that sued Proctorio and that I've lost my job as a librarian. I wonder how much his integrity cost.

I'll never forget my dinner with Brewster Kahle (and other brilliant folks) at the Internet Archive Canada headquarters. I am so inspired.

I found proof that Proctorio is being inflicted upon vulnerable children as early as kindergarten and talked to journalist Mark Keierleber about it. The story was published in The Guardian and The74 today. www.the74million.org/article/room...

Room Scans & Eye Detectors: Robocops are Watching Your Kids Take Online Examswww.the74million.org Remote proctoring tools like Proctorio have faced widespread pushback at colleges. Less scrutiny and awareness exists on their use in K-12 schools.

Reposted by

Ian Linkletter

Repeatedly encountering profs suggesting that chatgpt can help students “research.” But what’s never engaged is the fact that people doing research on a topic where they have little experience (as often happens in college) are the worst served by a tech that makes up information and citations.

Hey ChatGPT, why is Proctorio harmful? Answer: chat.openai.com/share/f6e390...

ChatGPTchat.openai.com A conversational AI system that listens, learns, and challenges

According to Turnitin's own press releases, the rate of papers being scanned by their AI detector has decreased from about 639,640 per day to 426,471 per day. They claim a "majority" of customers opted to license their AI detector after the free trial period, but use is down 33%.

Turnitin previously claimed that between April 2023 and November 9th, 2023, over 142 million papers had been scanned for AI. Now as of March 24, 2024, over 200 million papers have been scanned. What this means is that since November 9th, use has dropped 33%. That's the story.

In a press release, Turnitin says they have scanned over 200 million papers with its AI detector. They claim 3% of papers have 80% or more AI writing. They leave out how the 4% false positive rate they previously admitted to is harming millions of students. www.prnewswire.com/news-release...

Turnitin marks one year anniversary of its AI writing detector with millions of papers reviewed globallywww.prnewswire.com /PRNewswire/ -- Today, Turnitin celebrates the first anniversary of its award-winning AI writing detection feature. Since the launch of the tool in April 2023,...

And I'm the one being sued? Shaking my head... www.theverge.com/2024/4/6/241...

OpenAI transcribed over a million hours of YouTube videos to train GPT-4www.theverge.com How OpenAI, Google, and Meta deal with the limits of data online.

OpenAI removes ChatGPT's login requirement. This will have major implications on use, especially in schools. openai.com/blog/start-u...

Big ed tech news out of Ontario today. The Information and Privacy Commissioner released a report about Respondus Monitor and found that the company's use of student data to "improve the services" IS NOT AUTHORIZED BY FIPPA. Teresa Scassa's analysis is essential: www.teresascassa.ca/index.php?op...

Copilot is now banning certain numbers.

In response to complaints about harmful image output from Copilot, certain terms are now being blocked. Perhaps just as interesting is that there is a new warning that multiple "policy violations" will lead to suspension. This prevents whistleblowers from proving harm. www.cnbc.com/2024/03/08/m...

Microsoft begins blocking some terms that caused its AI tool to create violent, sexual imageswww.cnbc.com Microsoft has made changes to its AI guardrails after a staff AI engineer wrote to the FTC of his concerns about Copilot's image generation.

Reposted by

Ian Linkletter

NEW: We tested OpenAI’s GPT for racial bias systematically—and found it. Gift link: bloom.bg/3It4cE4

OpenAI’s GPT Is a Recruiter’s Dream Tool. Tests Show There’s Racial Biasbloom.bg Recruiters are eager to use generative AI, but a Bloomberg experiment found bias against job candidates based on their names alone

Reposted by

Ian Linkletter

Anthropic now admits using personal data. they updated their policy yesterday, which is now basically the same as OpenAIs