No joke. One of my students asked: Is AI the emergence of a conscious species? The road is long.
I was approached today by a journalist from a news agency, asking for comments on a "new scientific study challenging key indicators used in the last IPCC report regarding ocean heat uptake and the Earth's energy imbalance".
1/...
L'embauche de journalistes capables d'évaluer l'expertise l'est également. Cela éviterait de lire tout un tas d'inepties qui, par ailleurs, participent largement à la désinformation scientifique. Dans le domaine de l'IA par exemple, cela contribue à répandre des arguments marketing non fondés.
“But the AI tools powering translation are not always representative of how native speakers speak, said Aliya Bhatia, a policy analyst at the Center for Democracy and Technology who researches multilingual AI.”
"AI" is not a stochastic parrot.🦜
I wrote this piece a couple weeks ago, but it was hard for me to finish up given AI's role in society and war over the past few weeks. I should share it at some point though. Not perfect, but here it is.
medium.com/@margarmitch...
Registration are open until March 30th. The Grand Challenge on Multimedia Verification is open to anyone interested in verifying the authenticity and context of multimedia content: practitioners, students, researchers... @nordishub.bsky.social @edmo-eu.bsky.social
sites.google.com/view/mv2026/...
The Challenge is organised by Duc-Tien Dang-Nguyen (University of Bergen) Morten Langfeldt Dahlback (Faktisk), Anh-Duy Tran (KU Leuven), Minh-Triet Tran (University of Science, VNU-HCM), Chris Bregler (Google DeepMind). sites.google.com/view/mv2026/... 3/4
Participants will first explore 50 known cases to familiarise with the tasks. In the Validation stage, they must analyse 10 new cases and pass evaluation to qualify for the Real-World Verification stage, where 10 live verification cases will test their methods. 2/4
Intersted in #factchecking? The 2nd Grand Challenge on Multimedia Verification is open to everyone who wants to compete in verifying the authenticity and context of multimedia content. The competition consists of three stages: Training, Validation, and Real-World Verification. 1/4
Journalists have been asking us again and again about how to spot deepfakes.
Here’s our best advice for now, though it may evolve quickly, as AI advancements are happening in weeks and months rather than years. deepfakestracker.org/deepfake-spo...
Je viens de participer à un webinaire "L’IA redéfinit-elle les enjeux de la désinformation ?" organisé par le Conseil supérieur de l'éducation aux médias et EDMO Belux. Mon exposé portait sur l'IA racontée par la tech, l'occasion de développer cela ici : ohmybox.info/analyse-lia-...
Here is an example: "We already know that neural networks far exceed human capabilities (...) The problem is that AI has no intrinsic purpose" This is misleading because it conflates the specific task performance of NN with general intelligence, and it treats the lack of intrinsic purpose as a flaw.
And this one? I think these are two important moments in NLP history.
ieeexplore.ieee.org/abstract/doc...
If you are interested in doing high-impact, public-facing research ... Join Our Research Team! @SMLabTO is Hiring a Postdoctoral Research Fellow socialmedialab.ca/2026/03/04/p...
I have just started a new research project at the University of Bergen as part of the MSCA COFUND LEAD AI programme. Entitled 'F.A.C.T.' (Framing Authority and Critical Transparency in AI-Mediated Journalism), it will focus on AI-augmented search engines and their impact on journalism.
Among unethical reviewers are also those who expect authors to reshape a paper around how they themselves would have designed the study and who suggest several of their own papers as "essential" citations. This is completely unfair, but it sometimes happens.
Thousands of CEOs just admitted AI had no impact on employment or productivity—and it has economists resurrecting a paradox from 40 years ago fortune.com/2026/02/17/a...
I will always be amazed (and disappointed) by the way journalists treat all experts as if they were equal. They make no distinction between scientific experts whose research focuses on AI and self-proclaimed experts who spout nonsense...
Grammarly's new "Expert Review" tool uses AI to simulate feedback from famous authors and academics, including those who died decades ago. “These are not expert reviews, because there are no ‘experts’ involved in producing them,” says historian @ceaubin.com. www.wired.com/story/gramma...
#OutNow in #iCS
Dierickx et al examine how generative AI challenges established categories of factuality. Using fact-checking as an epistemological lens, they introduce the concept of “emergent facts” to assess AI-generated claims.
doi.org/10.1080/1369...
“…according to a report from Sweden's Svenska Dagbladet released last week. Employees in Kenya doing AI "annotation" told the journalists that they've seen people nude, using the toilet and engaging in sexual activity, along with credit card numbers and other sensitive information.”
🔥 New preprint 🔥
Guidance over guidelines? Unpacking the uses and concerns of generative AI in Communication Science 👇
osf.io/preprints/so...
WSJ reporting that the U.S. used Claude for the air strikes in Iran. Centcom has been using Claude "for intelligence assessments, target identification and simulating battle scenarios" www.wsj.com/livecoverage...
NYU is hiring two postdocs at the Center for Social Media, AI & Politics:
csmapnyu.org/jobs
And a grant manager in Sociology with the Social Science Research Hub:
uscareers-nyu.icims.com/jobs/15327/g...
I'm part of both groups--please share with anyone who is interested!
General purpose tools like ChatGPT, Claude, and Grok are not designed for this use, making mental health professionals wary.
Thank you so much!
I am excited to announce that on 2 March I will begin a MSCA COFUND Fellowship at the University of Bergen as part of the LEAD AI programme in the Department of Information Science and Media Studies. My research project focuses on search engines as AI-mediated environments in journalism.
A small number of samples can poison LLMs of any size www.anthropic.com/research/sma...
I was aware of the privacy issues on LinkedIn, but I didn't think it was this so serious and that the platform presented so many potential violations of the GDPR and the AI Act. thelocalstack.eu/posts/linked...
