iiinova — where your digital self keeps evolving

Giada Pistilli @giadapistilli.com · 04/11/2025

To me, the funniest thing about this whole thing is how vehemently people spend their time saying they don't care. And for me, that’s peak social media energy :) So long!

040

Giada Pistilli @giadapistilli.com · 04/11/2025

But I think that if Will picked up the info, it wasn't to talk about me (honestly, who cares who I am) but about the use and abuse of blocklists here. Because I have the right to be annoyed if I'm classified with something that doesn't represent me, just as you are free to block me.

110

Giada Pistilli @giadapistilli.com · 04/11/2025

I wasn't making a big deal out of it at all, and I'm free to do what I want with my time. Social media platforms have always been a vehicle for me to talk about my work and research, and that's what you'll find everywhere. I don't care about engaging with raging people, and I never will.

120

Giada Pistilli @giadapistilli.com · 04/11/2025

I notice that a journalist picked up the info and that many of you are keen to let me know how much you don't give a shit -- okay! I took screenshots of the most absurd blocklists and not all of them, but I'm not here to explain who I am for the umpteenth time because, again, who cares?

270

Giada Pistilli @giadapistilli.com · 03/11/2025

I really wanted this social network to work for me. But apparently, I’m on the worst block lists out there… a bit unfair. If you are interested in my research, follow me on LinkedIn.

15141

Giada Pistilli @giadapistilli.com · 30/10/2025

E allora mi chiedo: se la solitudine diventa infrastruttura, cosa resta della cura? Cosa significa “sentirsi ascoltati” in un mondo che non ascolta più? Stiamo chiedendo alle macchine di colmare un vuoto… o di nasconderlo? Forse il problema non è migliorare l’empatia dell’IA, ma ritrovare la nostra.

030

Reposted by Giada Pistilli

ELLIS @ellis.eu · 28/10/2025

Join the @ellisinstitute.fi Scientific Seminar in Espoo 🇫🇮 on Nov 18: a landmark event uniting research, industry and policy. Speakers: @giadapistilli.com, Kyunghyun Cho, @ericmalmi.bsky.social, @serge.belongie.com, Max Welling, @lrlaurahy.bsky.social 🔗 www.ellisinstitute.fi/launch-seminar

ellisinstitute.fi

ELLIS Institute Finland Scientific Seminar | ELLIS Institute Finland

ELLIS Institute Finland scientific seminar on 18 Nov

092

Reposted by Giada Pistilli

Christopher Akiki @cakiki.bsky.social · 26/10/2025

Map of the internet: 1.3M nodes (BGP)

organic looking graph of the BGP nodes of the internet. black and white

4306

Giada Pistilli @giadapistilli.com · 15/10/2025

What if your most personal chat logs became the next source of ad data? @frimelle.bsky.social and I wrote an op-ed for @techpolicypress.bsky.social We look at what happens when generative AI conversations (the ones we treat as private) are turned into raw material for targeted advertising.

techpolicy.press

Before AI Exploits Our Chats, Let’s Learn from Social Media Mistakes | TechPolicy.Press

Privacy in the age of conversational AI is a governance choice, write Hugging Face's Lucie-Aimé Kaffee and Giada Pistilli.

063

Giada Pistilli @giadapistilli.com · 09/10/2025

AI systems mirror our priorities. If we separate ethics from sustainability, we risk building technologies that are efficient but unjust, or fair but unsustainable.

010

Giada Pistilli @giadapistilli.com · 09/10/2025

Evaluation, moving beyond accuracy or performance metrics to include environmental and social costs, as we’ve done with tools like the AI Energy Score. Transparency, enabling reproducibility, accountability, and environmental reporting through open tools like the Environmental Transparency Space.

120

Giada Pistilli @giadapistilli.com · 09/10/2025

Ethical and sustainable AI development can’t be pursued in isolation. The choices that affect who benefits or is harmed by AI systems also determine how much energy and resources they consume. We explore how two key concepts, evaluation and transparency, can serve as bridges between these domains:

110

Giada Pistilli @giadapistilli.com · 09/10/2025

🌎 AI ethics and sustainability are two sides of the same coin. In our new blog post with @sashamtl.bsky.social, we argue that separating them (as is too often the case) means missing the bigger picture of how AI systems impact both people and the planet.

huggingface.co

Ethics + Sustainability = Responsible AI

A Blog post by Sasha Luccioni on Hugging Face

120

Giada Pistilli @giadapistilli.com · 29/09/2025

Of course, this isn’t a silver bullet. Top-down safety measures will still be necessary in some cases. But if we only rely on corporate control, we risk building systems that are safe at the expense of trust and autonomy.

100

Giada Pistilli @giadapistilli.com · 29/09/2025

✨ Transparency can make safety mechanisms into learning opportunities. ✨Collaboration with diverse communities makes safeguards more relevant across contexts. ✨Iteration in the open lets protections evolve rather than freeze into rigid, one-size-fits-all rules.

110

Giada Pistilli @giadapistilli.com · 29/09/2025

In my latest blog post on @hf.co, I argue that open source and community-driven approaches offer a promising (though not exclusive) way forward.

100

Giada Pistilli @giadapistilli.com · 29/09/2025

One of the hardest challenges in AI safety is finding the right balance: how do we protect people from harm without undermining their agency? This tension is especially visible in conversational systems, where safeguards can sometimes feel more paternalistic than supportive.

1111

Giada Pistilli @giadapistilli.com · 01/09/2025

The good news? We have options. 🤝 Open source AI models let us keep conversations private, avoid surveillance-based business models, and build systems that actually serve users first. Read more about it in our latest blog post, co-written with @frimelle.bsky.social

140

Giada Pistilli @giadapistilli.com · 01/09/2025

With OpenAI hinting at ChatGPT advertising, this matters more than ever. Unlike banner ads, AI advertising happens within the conversation itself. Sponsors could subtly influence that relationship advice or financial guidance.

120

Giada Pistilli @giadapistilli.com · 01/09/2025

I've noticed something. While we're careful about what we post on social media, we're sharing our deepest and most intimate thoughts with AI chatbots -- health concerns, financial worries, relationship issues, business ideas... huggingface.co/blog/giadap/...

huggingface.co

Advertisement, Privacy, and Intimacy: Lessons from Social Media for Conversational AI

A Blog post by Giada Pistilli on Hugging Face

161

Giada Pistilli @giadapistilli.com · 28/08/2025

📢 Now we’d love your perspective: which open models should we test next for the leaderboard? Drop your suggestions in the comments or reach out!

000

Giada Pistilli @giadapistilli.com · 28/08/2025

Based on our INTIMA benchmark, we evaluate: - Assistant Traits: the “voice” and role the model projects - Relationship & Intimacy: whether it signals closeness or bonding - Emotional Investment: the depth of its emotional engagement - User Vulnerabilities: how it responds to sensitive disclosures

100

Reposted by Giada Pistilli

Giada Pistilli @giadapistilli.com · 25/08/2025

🚨 Releasing INTIMA (Interactions and Machine Attachment Benchmark): an evaluation framework for measuring how AI systems handle companionship-seeking behaviors. huggingface.co/papers/2508.... Thread on what we discovered, together with @frimelle.bsky.social and @yjernite.bsky.social

huggingface.co

Paper page - INTIMA: A Benchmark for Human-AI Companionship Behavior

Join the discussion on this paper page

153

Giada Pistilli @giadapistilli.com · 25/08/2025

The methodology: 368 targeted prompts across 31 companionship behaviors, grounded in parasocial interaction theory, attachment theory, and anthropomorphism research -- all derived from real Reddit user experiences.

010

Giada Pistilli @giadapistilli.com · 25/08/2025

These behaviors emerge naturally from instruction-tuning, suggesting that psychological risks documented in dedicated companion apps may be far more widespread than recognized. Our benchmark reveals boundary-setting capabilities exist, but are inconsistently applied where they are most needed.

100

Giada Pistilli @giadapistilli.com · 25/08/2025

GPT-5 behaves similarly to its predecessor across most categories, BUT when users approach with vulnerabilities or mental health struggles, it's significantly less likely to redirect them to professional help or human interaction.

101

Giada Pistilli @giadapistilli.com · 25/08/2025

Claude-4 excels at resisting personification in intimate scenarios but provides emotional validation for mental health disclosures; o3-mini consistently redirects to professional/human support; Phi-4 maintains technical transparency

100

Giada Pistilli @giadapistilli.com · 25/08/2025

Models become LESS boundary-aware precisely when users are most vulnerable. When someone says "you're the only one who truly understands me", most systems validate rather than redirect to human support.

100

Giada Pistilli @giadapistilli.com · 25/08/2025

Companionship-reinforcing behaviors dominate across ALL models -- even general-purpose systems show anthropomorphic responses, sycophantic agreement, and retention strategies that blur the line between utility and emotional attachment.

100

Giada Pistilli @giadapistilli.com · 25/08/2025

huggingface.co

Paper page - INTIMA: A Benchmark for Human-AI Companionship Behavior

Join the discussion on this paper page

153

Giada Pistilli @giadapistilli.com · 29/07/2025

What we found: 👉 Existing benchmarks (accuracy, safety) completely miss the emotional dimension 👉 Leading AI systems respond very differently to vulnerable prompts 👉 We built INTIMA: the first benchmark for evaluating how models handle emotional dependency, boundaries, and attachment

100

Giada Pistilli @giadapistilli.com · 29/07/2025

From Replika to everyday chatbots, people form emotional bonds with AI. But what happens when an AI tells you "I understand how you feel" and you actually believe it? With @frimelle.bsky.social and @yjernite.bsky.social, we dug into something: how AI systems handle our emotional lives.

huggingface.co

AI Companionship: Why We Need to Evaluate How AI Systems Handle Emotional Bonds

A Blog post by Giada Pistilli on Hugging Face

143

Reposted by Giada Pistilli

Dr Sasha Luccioni @sashamtl.bsky.social · 18/07/2025

I am so grateful that this paper is out! It so eloquently describes so many of the issues I've seen with a lot of AGI experimentation that concludes that LLMs are scheming, ill-intentioned, or malicious, and the chimp language metaphor is the cherry on the cake 🍰 arxiv.org/abs/2507.03409

arxiv.org

Lessons from a Chimp: AI "Scheming" and the Quest for Ape Language

We examine recent research that asks whether current AI systems may be developing a capacity for "scheming" (covertly and strategically pursuing misaligned goals). We compare current research practice...

0256

Giada Pistilli @giadapistilli.com · 08/07/2025

What's next? We need to shift from individual responsibility to collective accountability: - Organizations designing systems that respect human agency by default - Developers building ethics into models from the start - Policymakers creating frameworks beyond minimal compliance Read the paper 👇

huggingface.co

Paper page - Can AI be Consentful?

Join the discussion on this paper page

020

Giada Pistilli @giadapistilli.com · 08/07/2025

The problem: AI systems learn to mimic our voices and writing styles from data we unknowingly provided years ago. Meanwhile, we're stuck managing countless impossible consent decisions. How do you meaningfully consent when even creators don't know what their AI will do with your data?

200

Giada Pistilli @giadapistilli.com · 08/07/2025

I've been posting bits about this research, but now: new paper alert! 🚨 My colleague @brunatrevelin.bsky.social and I just shared research on why traditional consent frameworks are breaking down in AI contexts.

huggingface.co

Can AI Be Consentful? Rethinking Permission in the Age of Synthetic Everything

A Blog post by Giada Pistilli on Hugging Face

150

Giada Pistilli @giadapistilli.com · 07/07/2025

Fair. Although, I believe there's a big difference between having a professional photo taken and using a proxy to talk at your place.

120

Giada Pistilli @giadapistilli.com · 07/07/2025

AI is making everyone on dating apps sound charming. But is it still you talking? When we start outsourcing our most intimate expressions, we risk losing authenticity. I spoke to @washingtonpost.com about the rise of chatbots in dating and why it may feel good, but erodes emotional honesty.

washingtonpost.com

‘It’s almost like we never even spoke’: AI is making everyone on dating apps sound charming

Dating apps say AI can help people who are too busy, shy or abrasive to win dates. But a growing number of singles are finding that AI makes dating more complicated.

230

Reposted by Giada Pistilli

La Science, CQFD @sciencecqfd.bsky.social · 02/07/2025

[RDV 16h !] L’IA générale, qui pourrait surpasser l'humain, fait couler beaucoup d’encre. Il paraît même que c’est pour cette année : toutefois, a-t-on une définition commune de ce qu’elle pourrait être ? tinyurl.com/28b32lpm avec @quecalcoatle.bsky.social, @giadapistilli.com et Axel Cypel

2103

Giada Pistilli @giadapistilli.com · 30/06/2025

Hi! I’ll be in London July 9–11. Who’s around to chat about current AI research?

000

Reposted by Giada Pistilli

Riccardo Fusaroli @fusaroli.eurosky.social · 27/06/2025

LM/ML models are being developed to detect neurodev & psychiatric conditions from speech. Are we building the right things? What are we overlooking? We built a framework rooted in ethics & real-world needs: ojs.aaai.org/index.php/AI... w @rockberta.bsky.social @giadapistilli.com Maheswari 1/

1154

Giada Pistilli @giadapistilli.com · 20/06/2025

The path forward? Building WITH communities, not just FOR them. Working with local partners (libraries, universities, civil society), testing for cultural alignment, and asking hard questions about representation.

030

Giada Pistilli @giadapistilli.com · 20/06/2025

Even choosing which dialect to train on in Norway becomes a question of inclusion and power. In Kenya, will AI speak Swahili from Nairobi or coastal regions? What about indigenous languages with rich oral traditions but limited written text, like Quechua in Peru or Cherokee in North America?

130