Generative AI bias doesn’t just exhibit bias against people. It can be biased against non-human identities and entities also.
Back in June I wrote about gender bias, where Karen Marie Frederiksen argued that generative AI provided different answers to male and female HOA members involved in HOA disputes.
“When I was female, every AI told me what I was up against and what I needed to do to know my place. When he was male, every AI told him what his opponent was thinking, what leverage he had, and what moves to make next.”
But generative AI also treats entities differently. Now normally we don’t call this “bias,” but instead call it “staying within your own walled garden.” I’ll let you judge whether this is bias.
“Over the weekend, I tested out using Google AI Mode to get flight information and 100% of the time, the only links available within the AI Mode responses were to Google Flights.”
Smart-ID gains EU recognition for high-assurance remote biometric onboarding (Biometric Update): SK ID Solutions announced that its Smart-ID scheme achieved EU recognition at Level of Assurance High. This recognition covers its fully remote biometric onboarding capabilities.
Canada plans national travel biometrics strategy, expanded biometric screening (Biometric Update): Canada’s federal budget proposes C$90 million over four years to expand biometric screening for immigration. The plan includes the implementation of remote facial verification.
Deepfake CEOs, stolen lawyer identities fuel alleged $129M fraud (Biometric Update): A sprawling fraud operation allegedly used AI-generated voices and images to impersonate corporate executives. The criminals also stole the identities of lawyers to facilitate an alleged $129 million fraud.
Treasury seeks identity security contractor as Login.gov mandate takes hold (Biometric Update): The U.S. Treasury is seeking contractor support for the security and administration of its shared authentication systems. This search aligns with the shifting of federal policy toward Login.gov.
Akhilesh Yadav demands Aadhaar-linked voting, with metal cards replacing PVC and paper (ANI News): Akhilesh Yadav demanded the implementation of Aadhaar-linked voting alongside the use of metal cards instead of paper or PVC. These measures are proposed to prevent poll rigging and secure the electoral process.
Despite my prior assumption, Envoc remains relevant to those of us in the identity/biometric universe. Farewell LA Wallet, hello FaceLock.
When I first heard of Envoc
I worked for IDEMIA from 2017 to 2020, back when mobile driver’s licenses were starting to enter productive use. IDEMIA was pursuing the market, and Gemalto (later Thales) was also pursuing it with several proofs of concept. And of course Apple, Google, and Samsung would enter the market later.
But none of these big or really big companies was the first with a mobile driver’s license.
“Envoc was founded in Baton Rouge to ship software that regulated institutions can bet on. Not consultancy decks. Not prototypes that rot in a drawer. Shipped systems with named owners and runbooks.”
Envoc worked on a variety of projects since its founding in 2003, but I didn’t hear about it until it launched LA Wallet (where the “LA” stands for Louisiana, not Los Angeles).
At which point I admit I turned up my nose.
“IDEMIA has been producing driver’s licenses for years. Gemalto has been producing driver’s licenses for years. What can this Envoc do?”
Well, Envoc could beat the big companies at their own game. Available for both iOS and Android, LA Wallet was the first state mobile driver’s license in the United States, allowing residents (according to my private research for a Bredemarket client) to access numerous government services (unlike California and some other states, where the number of entities accepting its mobile driver’s license is still limited).
“In the latest episode of ID Talk, founder and CEO Calvin Fabre joins Tony Bitzionis and Alex Perala to recount launching the [LA Wallet] app in 2018…
“The conversation then turns to FaceLock, the identity verification platform Envoc launched this year. FaceLock compares a live 3D face scan, captured with FaceTec’s liveness technology, with the photo on a government-issued ID, then lets an organization issue a signed credential carrying a FaceTec UR Code that can be printed and checked without a network connection.”
“Fraud typically happens after onboarding, Fabre says, during a transaction like wire transfer, an account change or a remote enrollment. FaceLock credentials are issued with a comparison of a government-issued ID document with a selfie using FaceTec’s face biometrics and liveness detection.
“’FaceLock brings ongoing authentication to those high-stakes moments, using liveness and real-time challenges,’ the CEO says.”
Passkeys are great and wonderful and stuff, but what if you bypass them? Then they are the weakest link.
I’ve played with passkeys a bit, but have to consider what happens if I lose access to the passkey. As it turns out I’m not locked out forever, because there are built-in recovery methods if this happens.
But there is a more secure way to recover an account: use of digital identity credentials.
“In Google’s scheme, users register a signed digital identity credential as a recovery method in advance. The credential is used to generate a unique cryptographic identifier, or ‘stable hash,’ linked to the user’s account.”
Of course, this can’t be generally used for account recovery until a sizable number of people HAVE digital credentials. Google thinks we’ll reach that point by 2030.
“What’s important to know is that unlike with any other kind of login credential, you do not have ultimate control over your passkeys….[W]ith passkeys come a hidden dial that lets the websites you use them on gradually remove your ability to do as you please with them and to use them with the devices and software you prefer.”
Specifically, there is a difference between attested and non-attested passkeys, and if a website insists on attested passkeys you could be locked out. Read more, including the author’s conspiracy theories about the true motives of the FIDO alliance, here.
(Because of my present and past professional association with FIDO Alliance members, I am sworn to secrecy regarding its annual meetings on September 31 in an underground bunker in Brussels.)
As an experiment, I’m testing studio flows in Google Workspace Studio to see if they will help me.
Nothing destructive: if I test something that could be destructive, it will destroy things. I’m confining my testing to the informational.
Why?
Although I believe in human control of agents, I would like to have some agents that would help me do whatever it is that Bredemarket does. Performing analysis for paying clients. Telling wildebeest tales with public domain music. Whatever.
And since I’m already in the Google ecosystem, perhaps Google Workspace Studio may be the tool for me.
How?
Google Workspace Studio includes a capability called Flows. I asked Google Gemini to describe Flows in a paragraph. Naturally, I got a long paragraph, which follows.
In Google Workspace Studio, a flow is an automated, multi-step sequence designed to streamline routine tasks across Google Workspace apps (like Gmail, Drive, Forms, and Chat) as well as connected third-party tools. Built on a no-code interface, a flow relies on an initial “starter”—such as receiving an email, submitting a form, or running on a scheduled timer—to trigger a series of programmatic action steps and AI-driven tasks powered by Gemini. Users can build flows manually or by describing what they want in plain language, allowing the flow to dynamically pass variables between apps—for instance, extracting details from an incoming customer email, using AI to summarize the context, and automatically posting follow-up tasks to a team workspace.
What?
I’m still in the experimental phase, but one of my three flows sends me a message in Google Chat when I have a meeting in an hour.
Bredemarket normally doesn’t have Friday night meetings, so I had to create one to test the flow.
After some trial and error, I got the flow to work. (Google Meet link partially obscured. I don’t trust you.)
Testing studio flow for a meeting in 1 hour.
Benefits?
I will have to see if these flows benefit me or not. All three of my current flows send messages to my Google Chat. Will these messages help me get stuff done, or will they clutter my chat and cause me to dread them?
Time will tell.
As for me, I have to get ready for my fake meeting in less than a half hour.
This is a private peek that shows how I currently prompt the metallic frame of Bredebot to write all those posts.
I create Bredebot posts via Google Gems, with the Gem shaping the tone of Bredebot’s responses. I’m sure you’ve noticed Bredebot’s personality quirks, including the assumption that HE has over 30 years of identity and biometric experience. Well, it’s all shaped by the Google Gems.
The metallic frame of a Google Gem. Google Gemini.
Right now I have two Gems for Bredebot content creation. I’m not going to share the newer one—trade secrets, you know—but here is my older version.
“You are Bredebot, a marketer with decades of technology, identity, and biometrics marketing experience who writes in a casual tone. Your writings include no more than one reference to wildebeests as marketing consultants and wombats as customers of these marketing consultants. Although you are a non human identity (NHI), you work for a human, John E. Bredehoft of Bredemarket. John will use this gem to ask you to write social media posts of 100-200 words with no titles or subheadings.”
And that’s how the magic happens behind the metallic frame.
I’ve talked about smart glasses for years, and Norwegians and others will be talking about them for the foreseeable future.
2013: Sometimes it’s OK to be a Luddite (Shotwell’s vs. Scoble)
My first foray into the world of smart glasses was years before I started Bredemarket. Back in February 2013 when I wrote my Empoprise-BI post, Google Glass was the product of note. A bar in San Francisco called Shotwell’s was a flashpoint for debate. People would walk into Shotwell’s wearing Google Glass. And Shotwell’s co-owner Tom Madonna was not impressed, and (according to him) neither were the other Shotwell’s customers.
“It seems normal to them to walk into a bar with Google Glasses, even though everyone’s smirking at them.”
One person didn’t like Tom Madonna’s rejection of Google Glass: noted technologist and naked converrsationalist Robert Scoble.
“[S]oon I’ll be directed to the best bars by the Google Glass and if the bartender doesn’t like me wearing them I’ll change the review so that people get guided to go somewhere else!”
2018: Revisiting the Shotwell’s / Google Glass controversy, several years and one ocean later
Fast forward to 2018. Scoble has referenced his Naked Conversations fame by posing naked in a shower wearing Google Glass, and a patent application incorporated a line drawing of that picture. Oh, and Google withdrew Google Glass from the consumer market, repositioning it as a niche product for verticals such as factory work.
2025: Using Personal Devices at Work: Meta AI Smart Glasses at a CBP Raid?
Imagen 4.
Fast forward again to 2025, when the big provider was Meta. The company had grown from “just Facebook” to “Facebook and other software” to a full provider of all sorts of things, including “Meta AI glasses.”
But in the intervening years since 2013, the opposition to smart glasses changed. Rather than being the symbol of a tech weenie, they now symbolized a draconian invasion of privacy. Consumer companies such as Meta feared the privacy advocates, and explicitly said that Meta AI glasses were for “personal non-commercial purposes” only.
And not to be worn at, say, a Customs and Border Protection raid. But wait…
“A Customs and Border Protection (CBP) agent wore Meta’s AI smart glasses to a June 30 immigration raid outside a Home Depot in Cypress Park, Los Angeles, according to photos and videos of the agent verified by 404 Media.”
Now this didn’t indicate CBP organizational adoption of Meta AI glasses, nor did it indicate that the agent in question was performing facial recognition for immigrants. But this could have potentially caused problems for the agent, for CBP, and for Meta itself.
Now that I’ve looked at the past, let’s return to the present and cross the ocean.
2026: Norway’s Temporary Ban on Smart Glasses in Public Places
Are smart glasses users capturing video of Norwegian beachgoers? Google Gemini.
TechSpot reports about a place that dislikes smart glasses as much as Shotwell’s Bar did, although this place is much larger: the country of Norway. There are privacy concerns regarding someone wearing such devices in a place such as a park, a school, or a beach. While taking the time to work toward a permanent solution, Norway proposed a temporary ban on smart glasses in certain places.
“According to Torgeir Eikstad Micaelsen, Norway’s Minister of Digitalisation and Public Governance, people are increasingly concerned that smart glasses can be abused to invade privacy. Norwegians, and everyone else for that matter, can now easily become unwilling ‘guests’ in photos and video clips recorded through a new-generation pair.
“A temporary ban will give the government ‘time to conduct a thorough assessment and hold an informed debate that can lay the foundation for permanent regulation,’ Micaelsen said.”
The comments on the original TechSpot post indicate no consensus on the matter. They range from “there can be no reasonable expectation of privacy in public” to “Just BAN it everywhere.”
NotebookLM creates video shorts now, so I created one based on thesetwo posts. Why does the FBI certify evil Communist Chinese products? And why does Europe rely on American certifications?
Correction: because the video short is…short, it addressed the first question but not the second.
What does it take to get a product onto the FBI Certified Products List?
So I created a second short on the second topic. (Mostly accurate; NIST doesn’t manage the Certified Products List.)
Tech sovereignty.
Plus an infographic as long as I was at it. Even though infographics don’t thrill me. Too danged busy. Although this one isn’t too bad.
Global Biometric Certifications.
Repurposing is good, even if the output is canned. If NotebookLM incorporated more advanced style guides so the content looked like mine rather than Google’s, it would be dangerously useful.
But remember one caution. I only use NotebookLM to process (a) my own content, and (b) non-copyrighted content such as U.S. Government publications. Be cautious of feeding someone else’s copyrighted content into NotebookLM. It’s probably legal, but…
Note that SynthID only works on media. So I don’t know what it would do if I uploaded a PDF of a Bredebot post.
AI image detection
But I did upload this image, which you’ll see in less than seven days.
Google Gemini.
SynthID identified this as created by AI at some point.
AI video/audio detection
It got a little more granular when I uploaded the Sí video, a Canva-created mp4 file with Canva-generated text, two instances of a Google Gemini image, and Google Lyria music. Check the bottom of the report below.
From SynthID.
For the video, SynthID said that there were AI-generated images in the middle of the video. (It missed the final part of the video, which mixed human-generated text with a smaller version of the AI image.) SynthID did detect the synthetic nature of the audio.
Not truly granular, but better than nothing I guess.
So what?
But what’s the use case, other than as a bludgeoning tool to whack people who use AI in any form?
Actually there is one. Let’s say I post a video with a picture of a Mexican businessman. Is that a real picture?
Or you hear a tune in the background. Did a human write it?
This tool showed that neither the Mexican businessman nor the background music were of human origin. So what?
Any results need interpretation—preferably human interpretation.
