Nick Diakopoulos
@ndiakopoulos.bsky.social
Northwestern University Professor. Computational Journalism; AI & Society. https://www.nickdiakopoulos.com. Writing https://www.ai-accountability-review.com/
“A.I. is a genuinely powerful technology that still needs people, and putting it to use is slow, human work. The sooner we are honest about that …, the sooner we can stop performing A.I. and start building something with it.” www.nytimes.com/2026/08/03/o...
Opinion | I Helped Run Lululemon. The A.I. Revolution Is a Hot Mess.
The A.I. revolution is stalling because companies don’t want to admit that integrating the technology is expensive and slow and requires human effort.
nytimes.com
New law in NY “requires social media companies to restrict algorithmically personalized feeds and nighttime notifications for users under the age of 18 unless they obtain parental consent” www.governor.ny.gov/news/governo...
Governor Hochul and Attorney General James Announce Final SAFE for Kids Act Rules to Protect Children Online
Governor Hochul and Attorney General James announced the release of final rules on how social media companies must restrict addictive features on their platforms to comply with the Stop Addictive Feed...
governor.ny.gov
Are your AI guidelines any good? New research published at @facct.bsky.social develops an assessment of quality: www.ai-accountability-review.com/p/are-your-a...
Are Your AI Guidelines Any Good?
New research on guidelines for evaluating AI guidelines
ai-accountability-review.com
I’ve been finding Claude Cowork to be really helpful for screening literature. Here’s my workflow: www.ai-accountability-review.com/p/claude-cow...
Claude Cowork for Research Screening
You don’t need Claude Science to support literature reviews
ai-accountability-review.com
A new faculty position in media and politics at Stanford: facultypositions.stanford.edu/jobs/assista...
Assistant Professor in Communication - Stanford University, California, United States
The Department of Communication at Stanford University seeks applicants for a junior appointment at the rank of tenure-track Assistant Professor. We invite applications from individuals working at the...
facultypositions.stanford.edu
A recent analysis of 188 autonomous AI incidents found that 35% involved code being destroyed or deleted. So as we delegate more to AI agents it's worth asking: How do we ensure people stay accountable for their agents?
I’m glad Anthropic models will be released, but de facto executive control of AI systems has got to be the worst possible version of AI governance. www.nytimes.com/2026/06/30/t...
U.S. Lifts Restrictions on Anthropic’s Most Powerful A.I. Models
nytimes.com
I'm collecting papers from #FAccT2026 @facct.bsky.social relevant to the AI Accountability Review (www.ai-accountability-review.com) in this thread. There are lots, but I'll post a few every day. 👇👇
AI Accountability Review | Nick Diakopoulos | Substack
The AI Accountability Review translates research to help bridge the gap between knowledge and practice for AI policymakers, practitioners, and researchers. Click to read AI Accountability Review, by N...
ai-accountability-review.com
I'm bummed to miss the FAccT Conference later this week in Montreal. If you're there please check out our work on simulating AI policy interventions which Julia Barnett will present (pre-print here: arxiv.org/abs/2605.27395). Another great collaboration with Kimon Kieslich and Natali Helberger.
Informing AI Policy Assessment using Large-Scale Simulation of Interventions
As the rapid proliferation of AI systems and harms spurs efforts in AI governance around the world, prioritizing among competing policy options has become increasingly challenging for policymakers and...
arxiv.org
What law exactly allows the president of the US to dictate that you can’t sell a product to a non-US citizen?
"Trump administration officials have asked a government artificial-intelligence testing unit to stop issuing public reports, the latest signal that the White House is tightening control over AI models as national-security concernsmoratorium?"
Exclusive | White House Reins In AI-Testing Unit as National-Security Concerns Grow
The unit’s future was thrown into doubt after its public work was halted despite winning praise from AI developers.
wsj.com
Court holds Google liable for false claims in AI Overviews. Seems significant. "A regular search engine just points to outside websites. But AI overviews generate 'independent, new, and substantive statements' by evaluating and combining content from various third-party sites" the-decoder
Landmark German ruling declares Google's AI Overviews are Google's own words and makes it liable for false answers
A German regional court has ruled that Google is directly liable for the content of its AI search overviews. According to the court, previous limited liability protections for search engine operators don't apply to AI overviews. In this case, Google's AI had falsely linked two publishers to fraud and made claims that didn't appear in any of the linked sources. The ruling could set a precedent for AI-generated content liability worldwide.
the-decoder.com
Interesting ideas from @hackshackers.com on making their site ready for AI agents: www.hackshackers.com/how-we-made-...
How we made our site more readable for AI agents and how you can too
Features for humans using AI agents and the AI agents themselves, including "ask" AI buttons and an MCP server
hackshackers.com
Today I updated my AI Accountability Review post on AI ethics principles: www.ai-accountability-review.com/p/ai-ethics-... — adding a new pointer to a great critical article on what’s needed to make AI principles more effective as governance from Raphael Max: link.springer.com/article/10.1...
AI Ethics Principles and Accountability
To move from high-level values to effective accountability, we still need to bridge the gap between abstractions and quantifiable, data-driven metrics.
ai-accountability-review.com
One of the best jobs in civic tech has just been advertised Head of AI at the mighty @fullfact.org
Head of AI - Full Fact
We’re hiring our next Head of AI. This is a rare and compelling opportunity, at a moment when our information environment is under significant strain.
fullfact.teamtailor.com
The number of journalists getting stung by this is just incredible. Like my mouth is on the floor that people just do not learn the lesson. Also book publishers.
Will Oremus@willoremus.com · 3mo agoIncredible www.nytimes.com/2026/05/19/b...
Today we're launching the Northwestern Agentic AI Investigative Journalism Challenge. Top prize is $5000 and an invitation to present the work at the Computation + Journalism Symposium this fall. Details here: www.gain-agent-challenge.northwestern.edu
The Agentic AI Investigative Journalism Challenge |
An Open Competition for Investigative TeamsThe Agentic AI Investigative Journalism ChallengeCoding agents can now work through messy document dumps, sur...
gain-agent-challenge.northwestern.edu
I am thrilled to be more involved with #C+J2026! Our CfP (deadline: June 29) centers "News Beyond the Newsroom". We welcome submissions that consider journalism's role in the broader and evolving information ecosystem. Please re-share broadly + let me know any Qs! www.cj2026.northwestern.edu/cfp/
Call For Participation | Computation + Journalism 2026 Symposium
cj2026.northwestern.edu
Nick Diakopoulos@ndiakopoulos.bsky.social · 3mo ago🚨 The Computation + Journalism Symposium is coming back to Evanston this fall! Get it on your calendar now for Oct 2-3, 2026. And participate: We’re soliciting contributions for research papers, workshops, and talks now through June 29. More info on the website here: www.cj2026.northwestern.edu
🚨 The Computation + Journalism Symposium is coming back to Evanston this fall! Get it on your calendar now for Oct 2-3, 2026. And participate: We’re soliciting contributions for research papers, workshops, and talks now through June 29. More info on the website here: www.cj2026.northwestern.edu
Homepage | Computation + Journalism 2026 Symposium
cj2026.northwestern.edu
Much of the world's population, including in the EU (based on our research), is asking for clear AI regulation. The EU today: Restrictions on high-risk uses of artificial intelligence in the EU will be postponed by more than a year! www.politico.eu/article/eu-c...
EU clinches deal to roll back AI restrictions
Deal marks first significant delay of digital rules amid pressure from the U.S.
politico.eu
Reward your model for being nerdy, get nerdy behavior. Train in your own output, pollute your dataset with verbal tics openai.com/index/where-...
Where the goblins came from
How goblin outputs spread in AI models: timeline, root cause, and fixes behind personality-driven quirks in GPT-5 behavior.
openai.com
Romance fraud has been accelerated by LLMs online. "LLM-powered scam agents achieve 46% victim compliance rates compared to 18% for human operators" link.springer.com/article/10.1...
Industrialized heartbreak: how generative AI enables romance fraud at scale - AI and Ethics
AI and Ethics - Romance fraud has undergone a structural transformation. What was once a labor-intensive confidence scheme, limited by the number of conversations a single operator could sustain,...
link.springer.com
In the future we’ll have a device, sort of like a continuous glucose monitoring sensor on your arm, that continuously verifies your identity via DNA while you use a computer. This will prove it’s you on the zoom and will stamp documents you create for authenticity.
Conversations with people in 2026: “are those the real numbers or did you get them from AI?”
How could AI explanations work to support accountability? I review some of the research here: www.ai-accountability-review.com/p/from-expla...
From Explanation to Accountability
A decade of explainable AI research has produced important techniques for understanding AI models, but less clarity on who those explanations are for and what accountability goals they actually serve
ai-accountability-review.com
Journalists should tread carefully when reporting on surveys these days -- the data could be based on a silicon sample, i.e. not from real people but from a simulation: www.nytimes.com/2026/04/06/o...
Opinion | It’s Called Silicon Sampling, and It’s Going to Ruin Public Opinion Polling
nytimes.com
If someone expects you to conjure up a new AI tool out of thin air, share these numbers for how much other newsrooms spent on their AI projects: generative-ai-newsroom.com/how-to-budge...
