Kate Kaye

@katekaye

Personal account. Researching algorithmic & surveillance tech in Portland, OR at RedtailMedia.org. Creator of Trust Control podcast. 25 yrs as a journo reporting on data use, AI, etc. Music nerd, forest walker, raptor watcher NY Mets fan. KEEPIN THAT PMA.

280 Siguiendo5051 Seguidores1197 Publicaciones

YouTube Bluesky

¿Es tu perfil?Reclámalo y katekaye.dirty.social es tuyo.Reclamar perfil

Kate Kaye@katekaye· 11d

"In a technical culture like Western culture, things you can't count, don't count." I reference the line from What Art Does by Brian Eno & Bette A. in my Trust Control documentary podcast, which ponders the human impulse to quantify & the blurring lines bw how computers trust & how humans trust.

Image shows the book "what art does" and two pages from the book, one with text and the other with an illustration of a bonsai tree

Kate Kaye@katekaye· 12d

Penis Description Fetish Ejaculation Dominant Hand These are data fields in Portland Police Bureau's Sexual Assault Management tracker system. -from slide presentation: www.sakitta.org/webinars/doc... -more info: www.govtech.com/analytics/Po... -related: www.oregonlive.com/crime/2024/1...

Image is a page from a digital slideshow titled "Peepers, Creepers & Exposers" that shows a list including terms such as Penis Description,
Fetish, Ejaculation and Dominant Hand.

Kate Kaye@katekaye· 22d

In my 2-part Trust Control podcast episode on media trustworthiness measurement I discuss Carl I. Hovland's WW2 US Army research studying effects of propaganda on recruits. It inspired his study of 'the sleeper effect' which said information lingers after distrust in the info source dissipates.

Kate Kaye@katekaye· 26d

Some of most influential research inspiring computational trust models (how computers 'measure' trust for AI agents & even how human worker trustworthiness is scored in relation to agents) was directly inspired by a 1930s method to measure 'aesthetic merit.' It ranked square #1, swastika #41, etc.

Kate Kaye@katekaye· Jul 13

Even the top US standards/measurement body, NIST, knows REDACTED medical data can be easily connected to public voter data to expose identity. It's been this way for decades. Add sophisticated AI to the mix and its identity exposure on steroids. www.nccoe.nist.gov/sites/defaul...

Kate Kaye@katekaye· Jul 7

As @runtime.news indicates, a new human trust scoring paradigm is emerging as AI agents creep into our work & personal lives. Humans (like open source devs) will be forced to prove humanity (even continuously). This is a thread throughout my TRUST CONTROL pod. Hear trailer: AITrustControl.com

An excerpt from Runtime.news reads: "The most important aspect of the Kubernetes policy is that a real human person must sign the code submitted and be prepared to answer questions about how it works. 'If you cannot personally explain changes that AI helped generate, your PR will be closed,' Hannon wrote on the Kubernetes blog."

Kate Kaye@katekaye· Jul 6

Speaking of Rand, US military complex has its fingerprints all over measure of trust in AI & automation. The inspiration for some of Trust Control's art is late '90s Air Force research out of U of Buffalo @ubuffalo.bsky.social - a scale of Trust in Automated Systems. Sign up at AITrustControl.com.

A 1998 University of Buffalo research paper shows a white page with black photocopied imagery featuring a globe with a grid. It says "UNITED STATES AIR FORCE RESEARCH LABORATORY -FOUNDATIONS FOR AN EMPIRICALLY DETERMINED
SCALE OF TRUST IN AUTOMATED SYSTEMS"

Kate Kaye@katekaye· Jul 6

Be sure to sign up to get notified when Trust Control launches (soon)! Visit AITrustControl.com and sign up at bottom right.

Kate Kaye@katekaye· Jul 5

Tucker's letter about Rand and Shapley: www.rasmusen.org/x/images/tuc... The original Stanford event Prisoner's Dilemma Game mimeograph handout: www.rasmusen.org/x/images/pd.... More info: rasmusen.org/GI/reader/no...

Kate Kaye@katekaye· Jul 5

Turns out Lloyd Shapley, the game theory guy whose work inspired SHAP was at a 1950 Stanford event where another game used throughout AI and automation trust measurement - the Rand corporation-designed "Prisoner's Dilemma" game - was first introduced under a different name. Links next.

Kate Kaye@katekaye· Jun 29

I haven't looked at all @jasonkoebler.bsky.social's photos from Cannes yet but this must be among my faves.

Kate Kaye@katekaye· Jun 8

Trust Control is coming. Sign up. aitrustcontrol.com A scripted podcast about: AI & Automation Trust Control at Work Scoring Trustworthiness of Media & Information Trust Scores for You and Your Agentic AI Co-worker The Military’s Role in Measuring Trust Trust Measurement History (and Residue)

A photograph of a vintage red rotary phone on a steel table. The handle is off the hook. There's a handwritten note on a square of white paper that reads, "Trust Control Trust is no longer a feeling."

Kate Kaye@katekaye· May 23

In my upcoming pod about the wacky world of 'trust measurement' I'll discuss residue of old trust models applied in questionable ways...like highly-cited '95 ABI model. Co-creator Roger Mayer talks in another pod about using it to measure HOW POLICE TRUST THE PUBLIC. www.youtube.com/watch?v=Sw0i...

Kate Kaye@katekaye· May 19

I went out early yesterday morning to Haystack Rock to peep some Puffins. Managed to capture a few snaps of them from very far away with my decent but obviously non-pro camera. My meager contributions to the Puffin fan club collection. I even got one of them snuggling!

A puffin sits on grassy moss high atop Haystack rock in Cannon Beach, Oregon. It is black with a white head and big orange beak and white hair-like feathers on its head.

Two puffins sit beside one another on grassy moss high atop Haystack rock in Cannon Beach, Oregon. They are black with white heads and big orange beaks and white hair-like feathers on their heads. One bird looks toward the camera and the other to the distance.

Two puffins cuddle close to one another on grassy moss high atop Haystack rock in Cannon Beach, Oregon. They are black with white heads and big orange beaks and white hair-like feathers on their heads. One bird snuggles its beak on top of the head of its companion.

Kate Kaye@katekaye· May 15

Mr. Met always brightens my day. (I know...so pathetic...but true.) Mets v Yanks this weekend: A big deal in my household of New York transplants on opposing sides! #LFGM

Kate Kaye@katekaye· May 13

In my podcast about Trust & Control, I'll discuss ways lines b/w human & AI "workers" are blurring. Corporations are calling AI agents "workers" alongside humans. They want continuous monitoring & verification of identity for everyday & WORK interactions. They say: Never trust. Always verify.

Kate Kaye@katekaye· May 11

A post suggested to me from a 'content producer' who was recently fired from a cloud tech company with a typically obtuse description. Ends the post stating, "In the meantime, let me know if your company needs help with human or even AI-assisted writing." Sucks he was fired, but, oh boy.

Kate Kaye@katekaye· May 9

For my forthcoming deeply-researched podcast series on "measuring" trust, reading this gem from Mckinsey, as if the governance-impedes-speed mantra is a new idea they're floating (From The agentic organization: Contours of the next paradigm for the AI era): Follow me for podcast launch alerts!

Highlighted text from McKinsey report states, "Ultimately, the scale of agentic adoption will be capped by how much oversight capacity humans can provide—making governance itself a potential bottleneck to productivity."

Kate Kaye@katekaye· May 7

Feeling hopeful about my New York Mets. Don’t leave me hangin on the telephone.

Kate Kaye@katekaye· May 4

In other news, the sky is blue.

Image is a news headline: "Data-Minimization principles and Training AI at Odds, Lawyer Says"

Kate Kaye@katekaye· Apr 28

I know academics are worried about AI generated papers but the amount of human-written fly-by-night journals out there is already nuts. I am not an academic and have had one paper published at one academic conference and reviewed papers at a couple others. I got this today:

A screenshot of an email says, "As an editor, I have had the opportunity to appreciate the significant contributions in your published work. In light of your esteemed expertise in the field, we would be honored to invite you to join our Editorial Board."

Kate Kaye@katekaye· Apr 23

It accurately associated me with an old book I wrote post '08 election about early days of voter data use and political ad targeting, and my work on algorithmic and surveillance tech use in Portland on my RedtailMedia.org site, which it called a 'platform.'

Kate Kaye@katekaye· Apr 23

4/5 When I asked it to explain itself & suggested that confusing two distinctly different names is a major error, it said it "built additional claims on top of that shaky assumption" and "turned a weak link into a misleading narrative." Ha, ya don't say.

ChatGPT wrote "What actually went wrong wasn’t just the names—it was a chain of issues: I asserted a specific paper connection without verifying authorship carefully enough
Then I built additional claims on top of that shaky assumption
And I framed her as more focused on kids/children’s tech than the evidence supports

That combination turned a weak link into a misleading narrative."

Kate Kaye@katekaye· Apr 23

3/5 Eventually, it admitted, "Despite earlier confusion Kate Kaye (the journalist) is not listed as an author on this paper. The paper is authored by Kate Mannell who is a different researcher (based @ Deakin University). The similarity in names (Kate Kaye vs Kate Mannell) can easily cause mix-ups."

ChatGPT wrote, "Despite earlier confusion, Kate Kaye (the journalist) is not listed as an author on this paper. The paper is authored by Kate Mannell, who is a different researcher (based at Deakin University).
The similarity in names (Kate Kaye vs. Kate Mannell) can easily cause mix-ups."

Kate Kaye@katekaye· Apr 23

1/5 Just gave ChatGPT a test to see what it might dig up about me and my journalism/research over the years. Some info was accurate, but it also told me a key area of my work was digital media impacts on kids. It told me I have done academic research on the topic. FALSE.

screenshot of chatgpt states "You’re right to call that out—that earlier connection was inaccurate, and your follow-up caught it. Without that check, it could have left a misleading impression about Kate Kaye’s body of work."

Kate Kaye@katekaye· Apr 4

Hey marketing team, can you come up with a fun name and image for our pay-with-your-iris-data system? Coming soon? Predictive analytics will recognize your iris and predict your shopping revenue value score!

An illustration of a person's face with large eyes connected to a large eye in the background. Text says "#BlinkLink PayEye scans your iris to make contactless payments SETUP NOW."

Kate Kaye@katekaye· Mar 30

System that uses eye movements, pupil response in attempt to detect drug/alcohol impairment is coming to law enforcement in EU (links in comments). Co also wants their tech in work settings. Here in US Dept of Transport has been testing this stuff for driving for years. www.nhtsa.gov/sites/nhtsa....

Kate Kaye@katekaye· Mar 23

The scene near Hanks Bar on N Robertson Street was super cool. Lots of 3D alligators, butterflies and fish. Unfortunately my photos of the awesome suit featuring a 3D smurf and colorful mushrooms didn't come out.

Kate Kaye@katekaye· Mar 23

Here are some shots from St. Joseph's Day (and night) in the Treme. The Black Masking Indians in the downtown area do a lot of incredible 3D style stuff that's distinct from the flat beaded patches of the uptowners. They're all amazing though.

Kate Kaye@katekaye· Mar 23

More Super Sunday uptown - I believe that's a Wild Man there with the skeleton mask and horns.