Can we detect #hatespeech at scale on social media? To answer this, we introduce 🤬HateDay🗓️, a global hate speech dataset representative of a day on Twitter. The answer: not really! Detection perf is low and overestimated by traditional eval methods https://t.co/OznvktglzK 🧵

ICYMI: Listen to @ManuelTonneau @oiioxford’s interview with the @SOEPTech podcast talking about his new research into hate speech, online platforms and disparities in content moderation across different European countries. More here:

Conversational AI is fast becoming a key information source for humans worldwide, including during election cycles. But what are the effects on users' epistemic health? 🧠 🚨Today we released new @AISecurityInst evidence that brings cautious optimism for LLMs vs internet search

@TheMediaLeader highlights new insights from @ManuelTonneau, @deeliu97, Prof. @ralphschroeder and Prof. @computermacgyve, whose research found that 16mn EU-based users of X “do not have moderators for their national language” — equivalent to 14% of the platform’s EU user base.

Millions of users are posting to social media and other platforms in languages with zero moderators, even within the EU. That's the topline finding from a working paper leveraging newly mandated transparency data under the DSA led by @ManuelTonneau

@oiioxford @JWI_Berlin @UCPH_Research @UMassAmherst @umsi @meedan

Finally tagging scholars whose work inspired this piece: @monaelswah @FarhanaShahid_ @NicProllochs @CatalinaGoanta @s_cresci Your feedback is most welcome!

This would also not have been possible without data collection efforts led by @JurgenPfeffer and without @claesdevreese @AUrman21 who made me aware of the DSA moderator count data on the other/better place 🦋 a while back

Had a blast working on this paper with my wonderful coauthors @deeliu97 @Antisomniac Kevin @ralphschroeder @EthanZ @computermacgyve

We also issue a recommendation: platforms and regulators should improve transparency by reporting moderator counts with context (eg content volume per language), ensure consistent reporting over time, and extend data coverage beyond EU languages.

So what? Main implication is that speakers of underserved langs likely receive less protection from online harms. Our analysis also nuances existing concerns: while Global South langs are consistently underserved, allocation for other non-English langs varies across platforms.

For languages with mods, we normalize mod counts by content volume per language and find that platforms allocate mod workforce disproportionately relative to content volume, with languages primarily spoken in the Global South (Spanish, Portuguese, Arabic) consistently underserved

We also quantify the amount of EU-based users whose national language does not have moderators, and we’re talking about millions of users posting in languages with zero moderators.

Taking Twitter/X as an example, we then show that languages subject to moderation blind spots are generally widely spoken on social media, representing an average of 31% of all tweets during a one-day period in countries where they are the official language.

We first look at language coverage and find that while larger platforms such as YouTube and Meta have moderators in most EU languages, smaller platforms (X, Snapchat) have several language blind spots with no human moderators, particularly in Southern, Eastern and Northern Europe

Concerns about underinvestment in non-English mod have long circulated via whistleblower @FrancesHaugen, but were never quantified. The EU’s Digital Services Act is a turning point, requiring platforms to disclose mod counts per language, making cross-lingual comparison possible.

Social media platforms operate globally, but do they allocate human moderation equitably across languages? Answer: no! -Millions of users post in languages with 0 mods -Where mods exist, mod count relative to content volume varies widely across langs https://t.co/VPfgRnKraM

Congratulations to OII researcher @ManuelTonneau for receiving an Outstanding Paper Award at #ACL2025! Manuel’s research explores hate speech detection models and their effectiveness for online content moderation.

Congratulations to @ManuelTonneau and co-authors on our paper receiving an Outstanding Paper Award @aclmeeting . Read it here:

🚨 New study! We tested whether AI-generated messages – single static messages vs. conversations – can boost intent to screen for colorectal cancer. Turns out: short, tailored AI messages outperform expert-written materials & match conversations, at a fraction of the time! 🧵👇

🚨 New preprint on AI persuasion and public health 🚨 A 3-min conversation with GPT-4o nudged HPV-vax-hesitant parents (who obv knew it was AI & consented!)—BUT reading standard public-health material still outperformed chatbots in impact and longevity. Details below 👇