Post

Visualizzazione dei post da febbraio, 2023

La comicità artificiale. Perché le AI non hanno il senso dell’umorismo? - Il Tascabile

Ragionare sulle AI ... significa anche ragionare sull’essenza stessa dei domini culturali a cui si applicano, testarne i limiti e riconfigurare la loro comprensione profonda. Vale per l’arte, vale per il giornalismo, per il copywriting, la letteratura, e vale anche per la comicità. Leggi tutto:  https://www.iltascabile.com/linguaggi/la-comicita-artificiale/

Le risposte di ChatGPT e delle altre intelligenze artificiali sono piene di pregiudizi - Wired

Da Google Translate a Character.AI fino a Chatsonic, i sistemi di deep learning addestrati per conversare non riescono fare a meno di ripetere i più classici stereotipi, nonostante le premure dei programmatori ... Dal momento che le intelligenze artificiali si limitano a rielaborare statisticamente i dati che vengono loro forniti, non potranno che riproporre nei contenuti da loro prodotti gli stessi pregiudizi insiti nella società. E così, mentre come società cerchiamo faticosamente di voltare pagina nei confronti di alcune discriminazioni e pregiudizi, i prodotti più avanzati della nostra epoca tecnologica – le intelligenze artificiali – rischiano invece di rafforzarli. Leggi tutto: https://www.wired.it/article/chatgpt-intelligenza-artificiale-testi-pregiudizi-bias/

Bing's ChatGPT-Powered Search Has a Misinformation Problem - Vice

Last Tuesday, Microsoft announced that its Bing search engine would be powered by AI in partnership with OpenAI, the parent company of the popular chatbot ChatGPT. However, people have quickly discovered that AI-powered search has a misinformation problem. An independent AI researcher named Dmitri Brerton wrote in a blog post that Bing made several mistakes during Microsoft’s public demo of the product. It often made up its own information, such as making up fake pros and cons for a pet vacuum, writing made-up descriptions of bars and restaurants, and reporting unfactual financial data in its responses. Read more:  https://www.vice.com/en/article/3ad3ey/bings-chatgpt-powered-search-has-a-misinformation-problem

AI Has Successfully Piloted a U.S. F-16 Fighter Jet, DARPA Says - Vice

On Monday, the US Department of Defense’s research agency, DARPA, announced that its AI algorithms can now control an actual F-16 in flight. The fighter aircraft that was first introduced in 1978 has now seemingly evolved into an autonomous plane. Read more:  https://www.vice.com/en/article/n7zakb/ai-has-successfully-piloted-a-us-f-16-fighter-jet-darpa-says

build a story, not a dashboard - storytelling with data

I had an impromptu consulting session yesterday with a friend who was looking for guidance on how to visualize data for an important new initiative he is spearheading at work. As he was animatedly sharing the context, anticipated resistance, and what he felt was a paramount opportunity, there was a term that surprised me, coming up in the conversation several times: dashboard. As in, the correctly constructed “dashboard” would solve everything. Read more:  https://www.storytellingwithdata.com/blog/build-a-story-not-a-dashboard

Scientists Made a Mind-Bending Discovery About How AI Actually Works - Vice

Researchers are starting to unravel one of the biggest mysteries behind the AI language models that power text and image generation tools like DALL-E and ChatGPT. Read more:  https://www.vice.com/en/article/4axjnm/scientists-made-discovery-about-how-ai-actually-works