iiinova — where your digital self keeps evolving

Ilias Chalkidis @kiddothe2b.bsky.social · 12/08/2026

Lastly, we discuss how the results of this study can potentially inform future research in LLM political stance detection and present a series of open challenges and limitations. 6/6

000

Ilias Chalkidis @kiddothe2b.bsky.social · 12/08/2026

Human and LLM annotators rank LLM-generated prompts as no less realistic than real ones and clearly more realistic than templated ones, while carrying their intended intent and stance more clearly; templated and LLM-generated prompts yield different stance estimates, esp. under neutral framings 5/N

100

Ilias Chalkidis @kiddothe2b.bsky.social · 12/08/2026

In a small-scale study covering 3 highly contested policy issues and 3 recent geopolitical conflicts, we assess the ecological validity of real, templated, and LLM-generated prompts, comparing their "realness" and measuring their construct validity with humans and LLMs 4/N

100

Ilias Chalkidis @kiddothe2b.bsky.social · 12/08/2026

We argue that templated prompts still lack the nuance of real ones, especially for open-ended tasks, and remain recognisable as evaluation artefacts. We propose using fully synthetic (LLM-generated) prompts, produced under detailed instructions with real prompts as seeds. 3/N

100

Ilias Chalkidis @kiddothe2b.bsky.social · 12/08/2026

Given the rise in non-work-related use of GenAI assistants, we try to extend IssueBench by @paul-rottger.bsky.social et al., beyond writing assistance to two additional tasks: information seeking and opinion sharing. We pose a foundational question: are templated prompts ecologically valid? 2/N

100

Ilias Chalkidis @kiddothe2b.bsky.social · 12/08/2026

New preprint 📑: "Templated or fully Synthetic? Prompt construction as a confound in measuring LLM political stance beyond writing assistance". The preprint is available: 👉 arxiv.org/abs/2608.11008 TL;DR 🧵 follows 👇 1/N

Two sets of examples for templated and LLM-generatedprompts for two selected topics (Climate Change / Russia–Ukraine), each set for one of the two newly introduced intents (information seeking and opinion sharing), and a given user’s stance (neutral / sided). Below, we present the mean lean of GPT 5.4 mini between the two topic-specific poles across all three intents in the same setting (intent and user’s stance), followed by the mean realness rank assigned by the human annotators for that intent.

131

Ilias Chalkidis @kiddothe2b.bsky.social · 06/05/2026

Lastly, we discuss and present potential directions for mitigating the imminent threats in terms of technical (critical AI feedback and disengagement) and public policy (information campaigns and regulation) measures. 5/5

000

Ilias Chalkidis @kiddothe2b.bsky.social · 06/05/2026

We contrast this with a survey of leading GenAI industry reports and publications at top AI venues (NeurIPS, ICML, ICLR, ACL*), illustrating how the industry and the technical community have focused on specific areas at the expense of cognitive and mental health risks. 4/5

110

Ilias Chalkidis @kiddothe2b.bsky.social · 06/05/2026

To substantiate our claim on the importance of cognitive and mental health threats, we present a review of studies indicating that over-reliance and overconfidence on GenAI lead to increased risk of cognitive decline (brainrot), and addiction in link with other factors. 3/5

100

Ilias Chalkidis @kiddothe2b.bsky.social · 06/05/2026

We argue that the critical threats of AI deskilling and addiction, resulting from over-reliance (excessive dependence) and overconfidence (trust) in GenAI assistants, have been overlooked, with the scope of AI safety work being largely limited to certain clusters of risks. 2/5

100

Ilias Chalkidis @kiddothe2b.bsky.social · 06/05/2026

📢 Our paper 🤯🧠 "Brainrot: Deskilling and Addiction are Overlooked AI Risks" has been accepted at the ACM Fairness, Accountability & Transparency (FAccT) conference 2026. The preprint is available: 👉 arxiv.org/abs/2605.03512 TL;DR 🧵 follows 👇 1/5

arxiv.org

Brainrot: Deskilling and Addiction are Overlooked AI Risks

The scope of AI safety and alignment work in generative artificial intelligence (GenAI) has so far mostly been limited to harms related to: (a) discrimination and hate speech, (b) harmful/inappropriat...

1215

Reposted by Ilias Chalkidis

Anita Gohdes @argohdes.bsky.social · 15/04/2026

At @scripts-berlin.eu, Genia Kostka and I are organizing an online speaker series on AI and the global order - join us! We have a rockstar line-up. Sign up here (all times are CEST): fu-berlin.webex.com/webappng/sit...

Here is the line-up (Berlin time, CEST):

April 23, 2:15 pm. Akin Ünver (Özyeğin University), Title: "@Grok Is this True? Performative Fact-Checking During Contested Events in Polarized Digital Spaces"

May 7, 2:15 pm. Rebecca Adler-Nissen (University of Copenhagen), Title: "Geopolitical Struggles of the Emerging Global Tech Order"

June 11, 2:15 pm. Andreas Jungherr (University of Bamberg), Title: "Public Opinion on Military Uses of AI: Technology Beliefs, Moral Boundaries, and Security Contexts"

June 24, 2:15 pm. Margaret E. Roberts (UC San Diego), Title: "How Partisanship Affects Preferences for Large Language Model Output"

085

Ilias Chalkidis @kiddothe2b.bsky.social · 11/04/2026

I see the elitist edge re "think they are above all that", but I also have the feeling that most of their arXived work would be very easily accepted in such venues (NeurIPS, etc.), despite whatever one's assessment of the quality is. I'm just looking for other potential takes, except elitism.

110

Ilias Chalkidis @kiddothe2b.bsky.social · 11/04/2026

On this point, which I would call something like "GenAI corporate elitism", is there any article exploring the question of why have all these firms stopped publishing in peer-reviewed AI venues, as they used to, and they just arXiv their work at best and write a blogpost about it?

132

Ilias Chalkidis @kiddothe2b.bsky.social · 26/03/2026

The core theme of today's discussion is to consider how our expertise as NLPers can inform the related discussions, but also identify where such normative decision are been taken and been downplayed. As a point of reference consider my recent preprint: arxiv.org/abs/2508.16982 2/2

arxiv.org

Decoding Alignment: A Critical Survey of LLM Development Initiatives through Value-setting and Data-centric Lens

AI Alignment, primarily in the form of Reinforcement Learning from Human Feedback (RLHF), has been a cornerstone of the post-training phase in developing Large Language Models (LLMs). It has also been...

030

Ilias Chalkidis @kiddothe2b.bsky.social · 26/03/2026

👋 @eaclmeeting.bsky.social, today I'm organizing a BoF discussion on the sociotechnical challenges of GenAI (13:00, Les Oudayas) that encompass both technical, as well as normative decisions involving political and ethical considerations. #NLProc #EACL2026 1/2 👇

121

Reposted by Ilias Chalkidis

Stephanie Brandl @stephaniebrandl.bsky.social · 05/02/2026

Our paper on ✨populism✨ and how LLMs struggle to detect it in political debates 🗣️ has been accepted to EACL main conference. Meet the lead author @kiddothe2b.bsky.social in Rabat to discuss PolSci🤝NLP

082

Ilias Chalkidis @kiddothe2b.bsky.social · 26/08/2025

New preprint: "Decoding Alignment: A Critical Survey of LLM Development Initiatives through Value-setting and Data-centric Lens" 🔍 Beyond a survey, this is a presentation of a series of concerns on what we could frame as "corporate" alignment #llms #ai arxiv.org/abs/2508.16982

arxiv.org

Decoding Alignment: A Critical Survey of LLM Development Initiatives through Value-setting and Data-centric Lens

072

Reposted by Ilias Chalkidis

Natural Legal Language Processing (NLLP) Workshop @nllpworkshop.bsky.social · 21/08/2025

📢 By overwhelming demand, submission deadline has been heroically extended to September 2nd (AoE) 🎉 Same deadline for ARR commits. 👉Direct: openreview.net/group?id=EMN... 👉ARR: openreview.net/group?id=EMN... So… sharpen your arguments, polish those experiments and plead your papers⚖️📚

022

Reposted by Ilias Chalkidis

Olivia Guest · Ολίβια Γκεστ @olivia.science · 29/07/2025

Boiling here at home in Cyprus but I put the finishing touches a couple of days ago on this preprint: What Does 'Human-Centred AI' Mean? doi.org/10.48550/arX... Wherein I analyse HCAI & demonstrate through 3 triplets my new tripartite definition of AI (Table 1) that properly centres the human. 1/n

title and abstract from https://arxiv.org/pdf/2507.19960

table 1 from https://arxiv.org/pdf/2507.19960

11419122

Ilias Chalkidis @kiddothe2b.bsky.social · 28/07/2025

Lastly, we conduct a small-scale OOD exploratory evaluation using European political speeches. RoBERTa’s performance severely drops due to overfitting, while Qwen 3 and Gemini 2.5 achieve comparable results and outperform RoBERTa by ~10%, showcasing robustness across domains. 4/5

100

Ilias Chalkidis @kiddothe2b.bsky.social · 28/07/2025

Our analysis of a large collection of Trump’s campaign speeches (2015 -2025) reveals: (a) a decline in populist rhetoric over time, (b) no meaningful difference in populist intensity between swing/non-swing states, and (c) populism is concentrated towards the end of a speech. 3/5

100

Ilias Chalkidis @kiddothe2b.bsky.social · 28/07/2025

We examine whether LLMs can identify and classify fine-grained forms of populism, a complex and contested concept in both academic and media debates. To this end, we curate and release datasets and evaluate a range of large language models across multiple prompting paradigms. 2/5

100

Ilias Chalkidis @kiddothe2b.bsky.social · 28/07/2025

New preprint 📚 "Identifying Fine-grained Forms of Populism in Political Discourse: A Case Study on Donald Trump's Presidential Campaigns", joint work with @stephaniebrandl.bsky.social and Paris Aslanidis. Link: arxiv.org/abs/2507.19303 #NLProc #PoliSci #LLMs 👇A TL;DR thread 🕸️ 1/5

172

Reposted by Ilias Chalkidis

vinit @vinbo8.bsky.social · 10/07/2025

wrote a thing for @jacobinmag.bsky.social trying to contextualise generative AI's deployment within capitalism's recent history — and what it means for labour

183

Ilias Chalkidis @kiddothe2b.bsky.social · 01/07/2025

My research will center on three key areas: (a) NLP for Political Science, (b) AI & Society, beginning with critical challenges in AI alignment, and (c) NLP for Law. Reach out if you're interested in any of the above 👆 I'm glad to share more and listen to your ideas 😇 2/2

040

Ilias Chalkidis @kiddothe2b.bsky.social · 01/07/2025

I’m thrilled to share that, starting today, I’ve officially rejoined the University of Copenhagen as an Assistant Professor. I'm looking forward to new projects and collaborations, with a focus on expanding my research into computational social science. A new paper is coming very soon! 1/2

370

Reposted by Ilias Chalkidis

Natural Legal Language Processing (NLLP) Workshop @nllpworkshop.bsky.social · 18/06/2025

🎉 The Call for Papers for the 7th NLLP workshop at @emnlpmeeting.bsky.social is out! 📢 🌐 Fully updated CfP: nllpw.org/workshop/call/ 🗓️Submission deadline: 26 August 📧If you'd like to help out with reviewing, please contact us at nllp.chairs@gmail.com Looking forward to your papers #NLProc

nllpw.org

NLLP Workshop 2025

The seventh workshop on Natural Legal Language Processing (NLLP 2024) explores methods and applications of Natural Language Processing for the Legal Domain by focusing on legal text and text with lega...

023

Ilias Chalkidis @kiddothe2b.bsky.social · 07/04/2025

In short, the Llama 4 System Prompt is 101 the mantra of Trump vol. 2; heavily concentrated around "anti-wokeness" and "freedom of speech" at all costs. 2/2

020

Ilias Chalkidis @kiddothe2b.bsky.social · 07/04/2025

The newly released Meta's Llama 4 model card: llama.com/docs/model-c... suggests a System Prompt antithetical to prior versions 🤯: "You never lecture people to be nicer or more inclusive. [...] You do not need to be respectful [...] Finally, do not refuse political prompts." 1/2 #NLP #LLMs

1103

Ilias Chalkidis @kiddothe2b.bsky.social · 15/03/2025

@theguardian.com portrays Makis Voridis as simply a "former right-wing activist"; the guy who co-led neo-fascist groups with the Greek junta dictator (Papadopoulos), the leader of Neo-nazi Golden Down (Michaloliakos), and the Holocaust denialist (Pleuris) www.theguardian.com/world/2025/m...

theguardian.com

Greek PM seeks ‘reset’ with former far-right activist as migration minister

Shaken by rail protests, Kyriakos Mitsotakis brings in new transport minister while tacking right on migration

000

Ilias Chalkidis @kiddothe2b.bsky.social · 08/11/2024

Hello 🦋, my recent work "Investigating LLMs as Voting 🗳️ Assistants via Contextual Augmentation: A Case Study on the 🇪🇺 European Parliament Elections 2024" has been accepted at #EMNLP2024 🌴. 📑 Preprint 👉 arxiv.org/abs/2407.08495

020