Skip to content
Notes by Axbom
Per Axbom Akkoma

Just printed all the explanatory models about AI I produced in 2023. Hadn't realised there were so many!

These are in Swedish but all have English counterparts as well.

The hammer has definitely been the most popular one. The most useful one is likely the communication model and/or the one outlining all the human actors in play.

Six A4 papers on a white tabletop, each with its own diagram, model or infographic.

🔗 Originally posted on Akkoma

9 Jan 2024, 20:52
Share
LinkedIn Bluesky Email
Per Axbom Akkoma

It's funny how my main issue with the reality of the Fast and Furious franchise is Dom's love of Corona. 😁

🔗 Originally posted on Akkoma

9 Jan 2024, 17:38
Share
LinkedIn Bluesky Email
Per Axbom Akkoma

Paper. Accidental Wiretaps: The Implications of False Positives By Always-Listening Devices For Privacy Law & Policy, by Lindsey Barrett and Ilaria Liccardi.

Abstract
Always-listening devices like smart speakers, smartphones, and other voice-activated technologies create enough privacy problems when working correctly. But these devices can also misinterpret what they hear, and thus accidentally record their surroundings without the consent of those they record, a phenomenon known as a ‘false positive.’ The privacy practices of device users add another complication: a recent study of individual privacy expectations regarding false positives by voice assistants depicts how people tend to carefully consider the privacy preferences of those closest to them when deciding whether to subject them to the risk of accidental recording, but often disregard the preferences of others. The failure of device owners to get consent from those around them is exacerbated by the accidental recordings, as it means that the companies collecting the recordings aren’t obtaining the consent to record their subjects that the Federal Wiretap Act, state wiretapping laws, and consumer protection laws require, as well as contravening the stringent privacy assurances that these companies generally provide. The laws governing surreptitious recordings also frequently rely on individual and societal expectations of privacy, which are warped by the justifiable resignation to privacy invasions that most people eventually acquire.

The result is a legal regime ill-adapted to always-listening devices, with companies frequently violating wiretapping and consumer protection laws, regulators failing to enforce them, and widespread privacy violations. Ubiquitous, accidental wiretaps in our homes, workplaces, and schools are just one more example of why consent-centric approaches cannot sufficiently protect our privacy, and policymakers must learn from those failures rather than doubling down on a failed model of privacy governance.

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3781867

#DigitalEthics

🔗 Originally posted on Akkoma

9 Jan 2024, 07:28
Share
LinkedIn Bluesky Email
Per Axbom Akkoma reply

@rkaj Right, how do I know I haven't caused the internal server error?

🔗 Originally posted on Akkoma

7 Jan 2024, 14:24
Share
LinkedIn Bluesky Email
Per Axbom Akkoma reply

@Mabande 😂 Tack för dagens bĂ€sta gapskratt.

🔗 Originally posted on Akkoma

7 Jan 2024, 11:44
Share
LinkedIn Bluesky Email
Per Axbom Akkoma

It's truly helpful how those error messages keep telling me it's an INTERNAL server error rather than just a server error.

Oh, I get it, that makes so much more sense. It's internal! Bloody hell, I thought someone was kicking the server with army boots every time I tried the website.

🔗 Originally posted on Akkoma

7 Jan 2024, 11:24
Share
LinkedIn Bluesky Email
Per Axbom Akkoma

Sometimes I get the feeling people wish I had a more positive outlook. They say that more people would take matters to heart if I instead focused on the many benefits of harmful tech. 🧐

Like sitting on a chair, my coffee on the table, with burning flames all around me and saying everything is fine.

That's certainly one way to go.

p.s. This is a jokey post. It's not that bad. As an optimist, I believe many things can be better. The way things get better is by first acknowledging what does not work well. Critique is a component of improving.

🔗 Originally posted on Akkoma

7 Jan 2024, 11:19
Share
LinkedIn Bluesky Email
Per Axbom Akkoma

Not surprised that the wise @molly0xfff is one of the people taking the lead in moving away from Substack.

She also made one of the best choices in my mind to go for an independent home, self-hosting Ghost. And managed to do it with minimum impact for existing subscribers.

I do know this is not for everyone, but also know that you don’t necessarily need a VPS to host Ghost, there is managed hosting available – and many more options than the quite expensive official one.

Molly White’s post on the hows and whys is here:
https://citationneeded.news/citation-needed-has-a-new-home/

🔗 Originally posted on Akkoma

5 Jan 2024, 21:28
Share
LinkedIn Bluesky Email
Per Axbom LinkedIn

Reading Dr. Joy Buolamwini’s Unmasking AI and I couldn’t recommend it more.

Love how she tells her own story and the difficult decisions she had to make to start addressing these harms that were apparent so early on for her. Even though she just wanted to code.

And the time when it turned out Cynthia Breazeal was in the front row, that honestly gave me goosebumps.

The chapter outlining how AI works is impressively accessible and ends on this important piece of wisdom: ”[
] if an automated decision impacts your opportunities and liberties, you must have a voice and a choice in whether and how technology is used.”

I mean, that pretty much sums up the challenges ahead. Read it and take it to heart. It’s really well-written, personal, engaging and brimming with wisdom.

https://www.unmasking.ai/

#DigitalEthics #AiEthics

🔗 Originally posted on LinkedIn

5 Jan 2024, 20:04
Share
LinkedIn Bluesky Email
Per Axbom Akkoma

Reading Joy Buolamwini’s Unmasking AI and I couldn’t recommend it more.

Love how she tells her own story and the difficult decisions she had to make to start addressing these harms that were apparent so early on for her. Even though she just wanted to code.

The chapter outlining how AI works is impressively accessible and ends on this important piece of wisdom: ”[
] if an automated decision impacts your opportunities and liberties, you must have a voice and a choice in whether and how technology is used.”

I mean, that pretty much sums up the challenges ahead. Read it, take it to heart. It’s really well-written and engaging.

https://www.unmasking.ai/

🔗 Originally posted on Akkoma

5 Jan 2024, 18:57
Share
LinkedIn Bluesky Email
Per Axbom Akkoma reply

@dangillmor

I'm feeling this too. Two years ago Substack defended prominent conspiracy theorists and anti-vaxxers profiting on the platform by saying that "the presence of writers with whom we strongly disagree” was a “necessary precondition for creating more trust in the information ecosystem as a whole.”
https://www.washingtonpost.com/technology/2022/01/27/substack-misinformation-anti-vaccine/

Nine months ago Nilay Patel got this response from CEO Chris Best on Decoder when asking if they would allow overt racism in Substack Notes: https://vimeo.com/817611949/a001a450c1

Substack haven't been hiding this stance. They've been enforcing it clearly, consistently and in the open for years.

🔗 Originally posted on Akkoma

5 Jan 2024, 17:41
Share
LinkedIn Bluesky Email
Per Axbom Akkoma reply

@caseynewton Did they ever change their stance on the conspiracy / anti-vaxxer people they were revealed to be profiting from two years ago? I feel like Substack have been saying exactly what they stand for, for quite some time now. Not sure how many strikes one considers them to be entitled to before they’re out


« This type of content is “so bad no one else will host it,” said Imran Ahmed, CEO of the Center for Countering Digital Hate, a nonprofit that focuses on combating misinformation and has researched Substack. By splitting subscription profits with creators, the group estimates, Substack generates at least $2.5 million a year in revenue from just five anti-vaccine leaders who have amassed tens of thousands of subscribers, each paying $50 a year.

Substack declined to comment, but shortly after The Washington Post made inquiries, CEO Chris Best and his two co-founders published a blog post saying that putting up with “the presence of writers with whom we strongly disagree” was a “necessary precondition for creating more trust in the information ecosystem as a whole.” »

https://www.washingtonpost.com/technology/2022/01/27/substack-misinformation-anti-vaccine/

🔗 Originally posted on Akkoma

3 Jan 2024, 20:20
Share
LinkedIn Bluesky Email
Per Axbom Akkoma

Substack has been profiting from conspiracy nuts and anti-vaxxers for many years. I feel like their stance on ”free speech” has been made clear. But it’s often the nazi angle that gets more people engaged.

I’m genuinely curious as to why Substack became such a well-respected platform when the writing was on the wall for so long.

I remember this controversy from January 2022 and Substack responded then exactly as they did now. Is it just that humans tend to forget so quickly? Obviously the fact that popular publications like Platformer approved of Substack, even after this revelation, makes people feel that it still comes recommended.

« This type of content is “so bad no one else will host it,” said Imran Ahmed, CEO of the Center for Countering Digital Hate, a nonprofit that focuses on combating misinformation and has researched Substack. By splitting subscription profits with creators, the group estimates, Substack generates at least $2.5 million a year in revenue from just five anti-vaccine leaders who have amassed tens of thousands of subscribers, each paying $50 a year.

Substack declined to comment, but shortly after The Washington Post made inquiries, CEO Chris Best and his two co-founders published a blog post saying that putting up with “the presence of writers with whom we strongly disagree” was a “necessary precondition for creating more trust in the information ecosystem as a whole.” »

https://www.washingtonpost.com/technology/2022/01/27/substack-misinformation-anti-vaccine/

Just remembered how Substack CEO Chris Best responded when Nilay Patel asked him on Decoder if he would allow overt racism on Substack Notes.

https://vimeo.com/817611949/a001a450c1

🔗 Originally posted on Akkoma

3 Jan 2024, 19:24
Share
LinkedIn Bluesky Email
Per Axbom Akkoma

49% of the world will be able to vote in an election this year. Good thing people have these new "AI" companions to help them get smarter about how to vote.

"One third of Bing Chat’s answers to election-related questions contained factual errors. Errors include wrong election dates, outdated candidates, or even invented controversies concerning candidates."

https://aiforensics.org/work/bing-chat-elections

🔗 Originally posted on Akkoma

3 Jan 2024, 14:06
Share
LinkedIn Bluesky Email
Per Axbom Akkoma

An absolute must-read by Erik Salvaggio on how image generators come to be and the abuse hidden within them.

”Stanford Internet Observatory’s David Thiel — building on crucial prior work by researchers including Dr. Abeba Birhane — recently confirmed more than 1,000 URLS containing verified Child Sexual Abuse Material (CSAM) is buried within LAION-5B, the training dataset for Stable Diffusion 1.5, an AI image tool that transformed photography and illustration in 2023. Stable Diffusion is an open source model, and it is a foundational component for thousands of the image generating tools found across apps and websites.”

”An additional point of serious concern is the likelihood that images of children who experienced traumatic abuse are influencing the appearance of children in the resulting model’s synthetic images, even when those generated images are not remotely sexual.”

« LAION’s data is gathered from the Web without supervision: there is no “human in the loop.” Some companies rely on underpaid labor to “clean” this dataset for use in image generation. Previous reporting has highlighted that these workers are frequently exposed to traumatic content, including images of violence and sexual abuse. This has been known for years. »

https://www.techpolicy.press/laion5b-stable-diffusion-and-the-original-sin-of-generative-ai/

🔗 Originally posted on Akkoma

2 Jan 2024, 17:26
Share
LinkedIn Bluesky Email
Per Axbom Akkoma

Grieving how lots of people with good ideas are opting for gated platforms for their content. It's hard to navigate all this, I get it, and I'm not blaming anyone but the platforms themselves.

Continuously hitting a wall on Medium is what got me this time. I just wanted to post a supportive comment...

🔗 Originally posted on Akkoma

2 Jan 2024, 17:04
Share
LinkedIn Bluesky Email
Per Axbom Akkoma

General-purpose search engines can lead you down all sorts of AI-generated delusions. While not a fail-safe, mastering DuckDuckGo's ability to bypass big search engines in favor of narrow ones can be helpful.

https://axbom.com/duckduckgo-bangs/

🔗 Originally posted on Akkoma

2 Jan 2024, 16:57
Share
LinkedIn Bluesky Email
Per Axbom Akkoma reply

@iamdavidobrien That I will do!! Lovely tips.

Here's I am holding a book at Waterstones on Princes Street :) It's funny, because it was actuallly our most recent trip to Edinburgh I was thinking of when I was considering where we probably bought the most books.

Happy new year David!

Hand holding the book Notes on a Nervous Planet by Matt Haig. Through the window in the background one can see Edinburgh Castle.

🔗 Originally posted on Akkoma

1 Jan 2024, 09:51
Share
LinkedIn Bluesky Email
Per Axbom LinkedIn

TÀnkte skriva nÄgot om skolan och digitalisering men jag har redan skrivit det. För snart sex Är sedan.

Det Àr ibland lÀtt att glömma att vi Àltar samma frÄgor om och om igen. Och att det vi skrivit och kommit fram till för mÄnga Är sedan fortfarande har vÀrde. Eller aldrig ens omhÀndertogs.

Jag la mycket möda bakom min text om mobilförbud. Jag har fÄtt mycket positiv Äterkoppling genom Ären (och de senaste dagarna).

Kanske har den ett vÀrde i att signalera hur man pÄ tydligare och tryggare sÀtt kan angripa frÄgor om digitalisering ur ett samhÀllsperspektiv.

Jag inbillar mig inte att du har tid att lÀsa den idag, pÄ nyÄrsafton. Den berÀknas ta nÀstan 15 minuter att lÀsa. Men kanske kan du hitta en stund under nyÄrsdagen att fundera över hur det kan bli om vuxenvÀrlden tar lite mer ansvar för sitt (vÄrt) agerande.

Planerar, utvÀrderar och resonerar. Inte bara sÀger vad vi ska göra utan ocksÄ tydligt talar om hur vi följer upp följderna av de experiment som barn och medborgare överlag utsÀtts för.

I stÀllet för att ideligen ignorera den djuplodande utvÀrderingen och hasta vidare till nya snabba lösningar som likt en hydra stÀndigt alstrar nya dilemman för nÄgon annan att hantera.

Och om du ansvarar för en publikation dÀr du tror att denna text eller nÄgot liknande skulle göra nytta... hör av dig.

Gott nytt omtĂ€nksamt Ă„r! ❀

Mobilförbud i skolan? Vi stÀller fel frÄga

🔗 Originally posted on LinkedIn

31 Dec 2023, 10:42
Share
LinkedIn Bluesky Email
Per Axbom Akkoma

How do you say "We have no idea how our product works", without saying those exact words.

OpenAI: "Training chat models is not a clean industrial process. different training runs even using the same datasets can produce models that are noticeably different in personality, writing style, refusal behavior, evaluation performance, and even political bias,"

They can say 'refusal behavior' and get away with it.

https://businessinsider.mx/chatgpt-accused-of-getting-lazier-2023-12/

🔗 Originally posted on Akkoma

31 Dec 2023, 09:32
Share
LinkedIn Bluesky Email
← Newer Older →
  • Notes by Axbom
  • RSS

© 2026 Notes by Axbom