Marielza
banner
marielza.bsky.social
Marielza
@marielza.bsky.social
Retired UNESCO Dir for Digital Inclusion, Policies & Transformation. Chair, UN University, eGov Institute.
UNESCO Women in STEM Committee

Some pottery and cyanotyping

Profile picture is of my face and torso

Banner is a picture I took of a light garden
Reposted by Marielza
I'm looking forward to the upcoming "Journal of Adversarial Poetry"
November 20, 2025 at 9:52 PM
Reposted by Marielza
Looks like LLMs are *very* vulnerable to attack via poetic allusion: "curated poetic prompts yielded high attack-success rates (ASR), with some providers exceeding 90% ..."

https://arxiv.org/html/2511.15304v1
November 20, 2025 at 5:06 PM
Reposted by Marielza
✨Exhibition: 
Our Shared World: Castle Mills Members' Show 5
Edinburgh Printmakers
28 November-15 March
Preview evening: 27 November, 6:30-8:30pm

https://f.mtr.cool/jnacsgdaeo

#cyanotype #printmaking #exhibition
November 20, 2025 at 1:02 PM
Reposted by Marielza
Old cyanotypes revisited 15
November 19, 2025 at 9:37 PM
Reposted by Marielza
Amen, Sarah from St. Louis.
November 20, 2025 at 6:20 PM
Reposted by Marielza
A ruling in Brazil’s Supreme Court reframes online speech in Brazil: stronger duties for platforms, but vague standards for “unequivocal knowledge” of illicit content. A win for accountability—or a risk to democratic debate? Joan Barata considers the outcome:
Brazil Supreme Court Ruling Redefines Framework for Platform Liability | TechPolicy.Press
Joan Barata considers the outcome of the Court's effort to address the liability of internet providers for content generated by third parties.
www.techpolicy.press
November 20, 2025 at 7:23 PM
Reposted by Marielza
Its near "super intelligence" "replacing professionals" in the PR rounds, then once everyone is convinced and uses them for legal and medical advice, following the deceptive marketing, they slip it in the terms of service that you shouldn't do that.

buttondown.com/maiht3k/arch...
OpenAI Tries to Shift Responsibility to Users
OpenAI is trying to shift the blame for bad legal and medical advice from its chatbot away from the company and onto users. We agree that no chatbot should be used for medical or legal advice.
buttondown.com
November 20, 2025 at 5:30 PM
Behold: a company, speaking on "the good of humanity."
🤮
“The Sam Altman-founded company behind the ‘Orb,’ a volleyball-sized metal sphere that scans irises to prove people's humanity, had a message for employees: work weekends and ignore anything outside your job — for the good of humanity.”
Sam Altman's eye-scanning Orb startup told workers not to care about anything outside work
Tools for Humanity CEO Alex Blania said that "if you should care about something else, and if you want something else, you should just not be here."
www.businessinsider.com
November 20, 2025 at 2:44 PM
Reposted by Marielza
As civic space collapses under surveillance, securitization, and shrinking aid, the tech world faces a choice—enable democratic participation or accelerate its erosion, writes Gina Romero, UN Special Rapporteur on freedom of assembly.
Civil Society Is At Risk—and Tech Is Part Of The Problem | TechPolicy.Press
The time for urgent, rights-based innovation in the aid and tech ecosystems is now, writes Gina Romero.
www.techpolicy.press
November 20, 2025 at 2:09 PM
Reposted by Marielza
On Wednesday, the Trump administration circulated a six-page draft executive order titled “Eliminating State Law Obstruction of National AI Policy.” It would create an “AI Litigation Task Force” led by the Attorney General to challenge state AI laws, among other provisions.
White House Circulates Draft Executive Order Targeting State AI Laws | TechPolicy.Press
The draft order, titled "Eliminating State Law Obstruction of National AI Policy," followed a Truth Social post by President Donald Trump.
www.techpolicy.press
November 20, 2025 at 2:10 PM
Reposted by Marielza
AI has entered elections, whether institutions are ready or not, writes Alberto Fernández Gibaja. He argues for a shared framework to guide when and how AI should be used in electoral processes, helping authorities strengthen resilience and protect democratic trust.
Why Electoral Authorities Need an AI Framework | TechPolicy.Press
Alberto Fernandez Gibaja discusses how electoral authorities worldwide view the challenge of AI in elections and recommendations for governance.
www.techpolicy.press
November 20, 2025 at 2:30 PM
Reposted by Marielza
*that signals Trump's desire to ban state AI laws
November 20, 2025 at 2:02 PM
Reposted by Marielza
"Yann LeCun, a pioneering scientist who led Meta’s artificial intelligence research efforts for more than a decade, is leaving the company to create his own start-up, he said on Wednesday."
Yann LeCun, a Pioneering A.I. Scientist, Leaves Meta
www.nytimes.com
November 20, 2025 at 4:35 AM
Reposted by Marielza
I'm editing an anthology about unlikable characters and seeking work from younger writers, 15-21. If you know of one such writer who might want to submit their work for consideration, please share this call for submissions: audacity.substack.com/p/acquired-t...
Acquired Tastes: A Call for Submissions
For younger writers, ages 15-21
audacity.substack.com
November 19, 2025 at 8:23 PM
Reposted by Marielza
I used to make cyanotypes myself (back before bread took over my life). Someday I will again.
November 18, 2025 at 8:30 PM
Reposted by Marielza
Continuing to muck about with cyanotype . Likely to start toning for different colors soon.
November 18, 2025 at 11:31 PM
Reposted by Marielza
The bookshelf in our beautiful seminar room @ksjatmit.bsky.social is filled with books by former fellows. Want to be next? Applications to our fellowship are open and on our website. @annaleen.bsky.social @ferrisjabr.bsky.social @tomzellerjr.com @s-r-m.bsky.social and Melanie Kaplan. Please share.
November 19, 2025 at 6:25 PM
Reposted by Marielza
Bruce Schneier on the Agentic AI trilemma: “Fast, smart, secure; pick any two. Fast and smart—you can’t verify your inputs. Smart and secure—you check everything, slowly, because AI itself can’t be used for this. Secure and fast—you’re stuck with models with intentionally limited capabilities.”
Agentic AI’s OODA Loop Problem - Schneier on Security
The OODA loop—for observe, orient, decide, act—is a framework to understand decision-making in adversarial situations. We apply the same framework to artificial intelligence agents, who have to make t...
www.schneier.com
November 19, 2025 at 3:40 PM
Reposted by Marielza
“GenerativeAI isn’t magic. It’s the product of millions of invisible creators—poets, coders, photographers, musicians—whose work has been harvested, scraped, & absorbed into AI models without consent or compensation…

This is not innovation. It is expropriation.”
—Otti Vogt
AI Must Fund the Commons – Not Eat Them
AI feeds on the cultural commons—now it must pay its dues, supporting the creators and public institutions it depends on.
good-leadership.beehiiv.com
November 15, 2025 at 12:23 PM
Reposted by Marielza
TikTok joins Pinterest in letting users throttle AI-generated content. This contrasts sharply with OpenAI and Meta, who are doubling down on 100% AI streams via Sora and Vibes.

The real challenge for traditional social apps is finding that sweet spot between human created content and AI slop.
TikTok will let you choose how much AI-generated content you want to see | TechCrunch
With the new AI-generated content control, users who want to see less of this sort of content can dial things down, while those who enjoy it can choose to see more of it.
techcrunch.com
November 19, 2025 at 7:32 PM
Reposted by Marielza
If you care about privacy - and these days you really should - then you definitely should be paying attention to the privacy of our neural data.
@wyden.senate.gov @daniellecitron.bsky.social @wbm312.bsky.social @gracecordovano.bsky.social @greatdismal.bsky.social

medium.com/@kimbellard/...
If You Could Read My Mind — Wait, You Can?
Over the years, one area of tech/health tech I have avoided writing about are brain-computer interfaces (B.C.I.). In part, it was because…
medium.com
November 19, 2025 at 3:55 PM
Reposted by Marielza
Reposted by Marielza
📣 Tenure-track DH position at Chapel Hill with a home in English/Comp Lit
Assistant Professor in Digital Humanities
The Department of English & Comparative Literature at University of North Carolina at Chapel Hill invites applications for a full-time Assistant Professor in Digital Humanities with an anticipated sta...
unc.peopleadmin.com
November 19, 2025 at 5:22 PM
Reposted by Marielza
For years we’ve repeated that “extreme poverty is rapidly falling.”

In a new analysis, @maxroser.bsky.social of @ourworldindata.org shows why that era is ending – and why hundreds of millions could remain trapped in deep poverty beyond 2030.

ourworldindata.org/end-progress...

#SDG1 #Poverty
The end of progress against extreme poverty?
In the last three decades, the world has made progress against extreme poverty faster than ever before. But unless the poorest economies start growing, this period of progress against the worst form o...
ourworldindata.org
November 19, 2025 at 3:12 PM
Reposted by Marielza
What the Commission calls “simplification” is in fact a sweeping rollback of core digital rights. Daniel Leufer of Access Now unpacks what the European Commission is really proposing — and what it means for digital rights in the EU and beyond.
Digital Rights Are on the Chopping Block in the European Commission’s Omnibus | TechPolicy.Press
Changes proposed under the Digital Omnibus point to a shift away from empowering people and towards granting discretion to business, writes Daniel Leufer.
www.techpolicy.press
November 19, 2025 at 3:28 PM