Henry Líu

📢 CFP: Cultural Diversity & Ethnic Minority Psychology Special Section: Open Science in Multicultural Psychology Empirical & conceptual work welcome across quant, qual & mixed methods 📅 LOI due May 15, 2026 🔗 tinyurl.com/CDEMP-OpenScie… #OpenScience #MulticulturalPsychology

The dominance of English-language data not only affects the answers chatbots give—it also shapes how they work. Fortunately, there are ways to improve the tech’s performance economist.com/science-and-te…

Even as they claim the right to train their models on work belonging to other people, the AI companies have rejected a similar premise when it comes to their own products, Alex Reisner reports: theatlantic.com/technology/202…

"If A.I. is capable of producing gripping fiction, and readers and editors are unable to discern its origins..." Then this article wouldn't exist? Bizarre to keep predicting this 'inevitability.' Once it happens, no one will care. But it hasn't happened! nytimes.com/2026/03/19/boo…

CfP: Journal of Non-professional Interpreting and Translation special issue on “Non-Professional Translation and Interpreting in Conflict and War: Power, Ideology, and Political Agency”. Abstract submission by June 15 2026 jonpit.unibo.it/announcement/v…

Great read! Confirms my hunch: Andreessen found a philosophy that validates his own shallowness and at the same time makes his AI investments look smarter. Adding a dash of neuro science to his business model. theverge.com/tldr/897566/ma…

The paper I’ve been most obsessed with lately is finally out: nbcnews.com/tech/tech-news…! Check out this beautiful plot: it shows how much LLMs distort human writing when making edits, compared to how humans would revise the same content. We take a dataset of human-written essays from 2021, before the release of ChatGPT. We compare how people revise draft v1 -> v2 given expert feedback, with how an LLM revises the same v1 given the same feedback. This enables a counterfactual comparison: how much does the LLM alter the essay compared to what the human was originally intending to write? We find LLMs consistently induce massive distortions, even changing the actual meaning and conclusions argued for.

New publication: 6th Edition Introducing Translation Studies Theories and Applications By Jeremy Munday, Sara Ramos Pinto, Jacob Blakesley routledge.com/Introducing-Tr…

'Sometimes I think about us ... as planets moving in certain directions, pulled by each other’s gravity.' Translator Sasha Dugdale on her collaboration with Maria Stepanova for our latest #BookClub pick, and the importance of holding onto our humanity: tinyurl.com/d5tbepy7

Join us next week for our next Thursday Seminar! 📣 Sofia Navarro Beck (Aarhus University) will be presenting: 'Behind the Screen: ‘A Case Study of a Danish Online Grooming Case’' 26th March 16:00-17:00 GMT Online only For more information visit: eventbrite.co.uk/e/behind-the-s…

Read an interview with Mathias Énard and me at @TheBookerPrizes : thebookerprizes.com/the-booker-lib…

“Literary translation is one of humanity’s most crucial tasks — one that should never be left to the machines.” Passionately argued piece on what we lose when we use AI to translate books. japantimes.co.jp/commentary/202…

⛑ #Bilingual consent forms significantly improve cancer treatment understanding among people with limited English, with understanding rising from 35% to 60%, a new study finds eurekalert.org/news-releases/…

The new #LinguapaxReview 2025 is now online 📰✨ This issue looks at the links between heritage languages and speakers’ physical, psychological and social well-being, with voices and case studies from around the world. 🌍 ▶️You can download it here: linguapax.org/en/publication…

In an age where LLMs provide a wealth of data and analysis upon request, possibly one of the best ways to learn about the world is to travel, and learn about global cultural differences.

It's good to see that people now understand that AI slop isn't just bad AI Gen art - it's slop law, slop policing, slop education, slop military. Everywhere AI is forced into systems it brings with it the same +30% error rate of hallucinated info. Slop is dangerous.

Create enough hallucinated legal arguments, flawed engineering calculations and backdoor-ridden code, and the slop vats fill faster than our capacity to tell good work from bad, writes Tim Harford.⁠ ⁠ Read his column on telling good AI from bad: ft.trib.al/j6Io85O

In the latest installation of his Translated Chinese Literature column, Jack Hargreaves introduces five new books that ask us what it means to be human: chinabooksreview.com/2026/03/17/tcl…

"What remains most emblematic [...] is the electrifying shock it sends to the reader’s system." Rebecca Suzuki reviews Kanako Nishi’s SAKURA—a bracing novel exploring love's complexities—that gains new life in Allison Markin Powell's translation: asymptotejournal.com/blog/2026/03/1…