S G Dickinson

S G Dickinson

@sgdickinson.bsky.social

Internet & digital policy, government use of ICTs, women in IT. I tweet live from relevant meetings.

296 Followers 281 Following 630 Posts Joined Oct 2024
5 days ago
Preview
Gig workers in Africa have been helping the US military. They had no idea Data labelers for the company Appen say they have little insight about how their work is used.

Gig workers in Africa helped train AI systems for a company that performed work for a secretive U.S. military unit. Many were never told their work could be used for military purposes https://restofworld.org/2026/gig-workers-us-military-appen/?utm_campaign=row-social&utm_source=bluesky

40 26 2 6
5 days ago
Instructions on how to load a live transcript back to the start of the transcript session:

If the base URL is "https://www.streamtext.net/player?event=NAME":

- Add " "/transcript" after "player" but before "?" to produce

https://www.streamtext.net/player/transcript?event=NAME

With the #ICANN85 GAC (Governmental Advisory Committee) making full use of a human-generated transcript, it avoids the many bizarre clangers that Zoom AI captions tend to generate, so well worth checking it out. A reminder on how to get back to the start of a live transcript session:

1 0 0 0
5 days ago
ICANN85 Mumbai Community Forum | ICANN Public Meetings Stay Informed and Engaged with the ICANN85 Community Forum in Mumbai – Learn about how to attend and register at the Official ICANN Meetings Site.

#ICANN85 is underway in Mumbai. I'm following remotely, but on a road trip, so my coverage of this final ICANN meeting before the launch of the next round of new gTLDs could be a bit patchy.

0 0 1 0
5 days ago

Some wanted the Global Dialogue on AI Governance at the Palais des Nations, where it would been hard for observers to get to, with separate UN badging needed & delegates focused on AI having to shuttle between the Dialogue (Palais) & WSIS Forum (Palexpo). Thankfully that scenario has been averted.

0 0 0 0
5 days ago
Roadmap | Global Dialogue on AI Governance

The location of the 1st Global Dialogue on AI Governance has been confirmed: Palexpo, Geneva, 6–7 July 2026, just before the Global AI for Good Summit & in parallel to the WSIS Forum. All three events will take place at the Palexpo.

0 0 1 0
1 week ago
Preview
Children, Technology and Education in Conflict - Security Council, 10113rd meeting Maintenance of international peace and securityChildren, technology and education in conflict

Webcast archive of today's Melania Trump-chaired United Nations Security Council session now available.

1 0 0 0
1 week ago
Preview
Anthropic’s AI model Claude gets popularity boost after US military feud Claude climbs to top of app store charts in US and UK after being blacklisted by Pentagon over ethics concerns

"Altman [...] said the military wouldn't use ChatGPT 4 autonomous killing systems or mass surveillance. But [there's] skepticism by many AI experts, lawyers, tech workers & users, who asked why the US govt would abandon [...] Anthropic, only to strike a deal with OpenAI that has the same safeguards"

1 1 0 0
1 week ago
Russia speaking at the United Nations Security Council session being chaired by Melania Trump, with one of their female delegates wearing a top with a pussy bow.

Russian delegation includes a pussy bow wearing woman at the UNSC session being chaired by Melania Trump. Coincidence?

0 0 0 0
1 week ago

Chaos ensued as no delegate was allowed to take phones or cameras into venue. My small rebellion was to use my laptop webcam to take photos.

All planned parallel sessions were cancelled.

Lots of very angry delegates unhappy a president's wife could coopt an int'l meeting to promote herself.

/2

0 0 0 0
1 week ago

Watching Mrs Trump chair the UNSC session on children & tech is giving me flashbacks to IGF 2009 in Sharm el-Sheikh, when Mrs Mubarak coopted the forum to hold her own last-minute session, "Preparing the Young Generations in the Digital Age: A Shared Responsibility". /1

0 0 1 0
1 week ago
Unilateral governance

But perhaps the most consequential—and most revealing—choice in the declaration is institutional in nature. The document strips the United Nations of central responsibility for AI governance and places its bets on the multi-stakeholder model, that is, a governance shared among states, companies, and other non-state actors, without any clear hierarchy of responsibilities and with decisions made by consensus. This is not a neutral choice: it is precisely the model that shaped internet governance over the past few decades, with results that historical experience itself allows us to assess.

ICANN, the internet’s principal technical coordination body, was born with the rhetoric of multi-stakeholderism and produced, in practice, a progressive corporate capture, marked by a lack of transparency and control by a single power. The transition from North American state oversight to this model did not represent a victory for globalism, but the replacement of state control with control by the United States’ major technology companies. New domains were sold at prohibitive prices; civil society was treated as a decorative layer; accountability mechanisms proved largely performative; and nation-states have no active voice in its decisions. The result was the consolidation of an internet whose logical, economic, and infrastructural layers remain under the unilateral control of a small group of countries and corporations. For AI, the risk is analogous and potentially more serious. The extremely high capital requirements for developing large language models make multi-stakeholderism even more fragile than it was in the case of the internet: without billions of dollars in processing power, data, and talent, there is simply no way to sit at the table as an equal. The model proposed by the declaration thus paves the way for bilateral governance between Big Tech and major states—with the Global South reduced to the role of consumer and data supplier. The so-called AI Summits risk repeating the trajectory of the WSIS—the World Summit on the Information Society—a process created in 2003 that dragged on for decades of rhetorical stalemate without any concrete results in terms of redistributing power over digital infrastructure.

The choice between multilateralism and multi-stakeholderism is not technical; it is political. Multilateralism, for all its flaws, anchors responsibilities in sovereign states and preserves the possibility of binding mechanisms negotiated under conditions of formal equality. Multi-stakeholderism, as the history of the internet has demonstrated, tends to dilute that formal equality into real asymmetries—where the weight of each actor depends less on its democratic mandate and more on its economic and technological power.

"Extremely high capital requirements for developing LLMs make [AI governance] multistakeholderism even more fragile than it was in the case of the internet: without billions of $ in processing power, data & talent, there's simply no way to sit at the table as an equal"

circleid.com/posts/ai-the...

2 0 0 0
1 week ago

I assume (hope) that when ICANN staff are formally back at work on Monday, Los Angeles time, someone with permissions to edit the website's homepage will update the #ICANN85 travel advice news item (currently linking to older 24 Feb item) to link to the most up-to-date announcement (28 Feb).

1 0 0 0
1 week ago
Screenshot of ICANN website - icann.org - homepage, advising users to ignore the big news item on travel avice for ICANN85, and choose the "More announcements" link, instead, to find the most recent travel advice on ICANN85.

For anyone flying to #ICANN85 in Mumbai & checking the ICANN homepage obsessively to see if there's updated travel advice but only seeing the big news item that predates the Iran bombing, there's actually updated advice. You just have to scroll down, click "More announcements", then "Travel Alert".

0 0 1 0
1 week ago
Preview
The Age Verification Trap Platforms are caught in a dilemma: enforce age restrictions or protect user privacy. How are they navigating this tricky terrain?

"The only way to prove someone is old enough 2 use a site is 2 collect personal data about who they r. The only way to prove u checked is to keep the data indefinitely. Age-restriction laws push platforms toward intrusive verification systems that often directly conflict w modern data-privacy law"

0 0 1 0
1 week ago

Wondering what effect bombing Iran will have on upcoming digital policy related mtgs:

CSTD WG on Data Governance, 2-3 March, Geneva
ICANN85, 7-12 March, Mumbai
ITU Council, 28 April-8 May, Geneva
ITU Plenipot, 9-27 November, Doha

a) Flights will be an issue for many
b) Existing tensions heightened

0 0 0 0
1 week ago
Post image Post image Post image

Iran has now been largely offline for 18 hours, with an #Internet shutdown put into place following strikes by the US & Israel. However, small amounts of Web & DNS traffic are still getting through, and IPv4 routes remain available, suggesting some users and sites are allowlisted for access.

19 17 0 0
2 weeks ago

Australian eSafety Commission study to follow 4000 kids+families for 2 years to see how the Social Media Minimum Age law's working (or not) & how it impacts kids.

Frankly, I think parental education is key. I know a dad of teen daughters who's perfectly OK with AI nudification apps being available🫤

0 0 0 0
2 weeks ago

Yet another reason to always keep humans in the loop ... "AIs can’t stop recommending nuclear strikes in war game simulations. Leading AIs from OpenAI, Anthropic and Google opted to use nuclear weapons in simulated war games in 95 per cent of cases" www.newscientist.com/article/2516...

17 18 2 3
2 weeks ago

We are very susceptible to AI manipulation.

"Even when AI tools provide source, people are far less likely to check it out than they were with old-school search results. For ex, a recent study found ppl are 58% ⏬️ likely to click on a link when an AI Overview shows up at the top of Google Search."

0 1 0 0
2 weeks ago

AI Impact Summit Declaration, New Delhi (Feb 18-19, 2026) is out.

As of 10 UTC, 22 Feb, "Endorsing countries/int'l organizations" are 87 countries & 1 UN agency: IFAD (Int'l Fund for Agricultural Development).

But you just know it's a bland doc when even the current US administration signs it.

1 1 0 0
2 weeks ago
Preview
How tech turned against women As AI-generated sexualised images proliferate and app-facilitated abuse spreads, we are sleepwalking into a new age of gender inequality. It is time to regulate properly

"large language models such as ChatGPT were consistently advising women to ask for lower salaries than men in recruitment processes,... AI tools already in use by more than half of England’s councils were downplaying women’s medical conditions, potentially resulting in unequal care"

957 522 23 68
2 weeks ago
Preview
Man receives £42,000 bill for data roaming charges after Morocco holiday Andrew Alty thought O2 bill was a mistake, but daughter’s TikTok use allowed massive uncapped charges to accrue

Interesting that it's the telco's uncapped roaming charges focused on here and not the fact that social media is designed to be so addictive that a kid preferred to be glued to her phone rather than enjoy the experience of a different country right in front of her.

0 0 0 0
2 weeks ago
Preview
Amazon’s cloud ‘hit by two outages caused by AI tools last year’ Reported issues at Amazon Web Services raise questions about firm’s use of artificial intelligence as it cuts staff

When you drink your own AI Kool-Aid across your critical infrastructure this can happen:

0 0 0 0
2 weeks ago

10/ Pre-Trump, other countries tolerated US dominance because the ideals US proselytized, if flawed in practice, were seen as generally morally/ethically desirable & the basis for a better world.

But today's US ideals are seen as harmful to all but a tiny minority. Not something the world wants!

0 0 0 0
2 weeks ago

9/ Whatever the Trump admin does with freedom[dot]gov, it's going to face the same problem. It's trying to control, unilaterally, what the rest of the world sees on the Net. And as we all know, single points of control are inevitably single points of failure...

0 0 1 0
2 weeks ago

8/ The Trump admin could invoke Musk's Starlink terminals to bypass any EU restrictions to freedom[dot]gov domains & IP addresses. Localized jamming is possible but can be bypassed. If so, ITU's radiocommunication sector could become the venue for the mother of all digital sovereignty battles.

0 0 1 0
2 weeks ago

7/ Instead, freedom[dot]gov focuses on channeling extremist "free speech" to like-minded, white-skinned MAGA types in EU countries, but could fail even in that narrow remit, based on the information currently available about the project.

0 0 1 0
2 weeks ago
Preview
US funding for global internet freedom ‘effectively gutted’ Programme that funds groups building tech to evade oppressive government controls under serious threat

6/ By centralizing everything via freedom[dot]gov, the US govt removes the advantages the old Internet Freedom program had in supporting grassroots, agile processes aimed at bypassing authoritarian regime's Internet restrictions.

0 0 1 0
2 weeks ago

5/ Savvy users will always find ways to bypass specific restrictions. A big part of that is due to the Net being decentralized & govts inevitably being slower at digital whack-a-mole than the combined global efforts of individuals who are agile & "creative" in finding ways to bypass local laws.

0 0 1 0
2 weeks ago

4/ If the Trump admin is planning to invoke TOR for freedom[dot]gov as a way to bypass governments enforcing their sovereign national laws to regulate content within their borders, that can be blocked, too.

0 0 1 0