#OpenAI just made its first #cybersecurity investment
https://techcrunch.com/2025/04/03/openai-just-made-its-first-cybersecurity-investment/
US authors’ copyright lawsuits against OpenAI and Microsoft combined in New York with newspaper actions
California cases over AI trainers’ use of work by writers including Ta-Nehisi Coates and Michael Chabon transferred to consolidate with New York suits from John Grisham and Jonathan Franzen and more
The UK wants to introduce an opt-out copyright system that would mean #AI firms can use anything for generative AI training unless explicitly forbidden. This idea is heavily opposed by creatives.
But it's still not permissive enough for #Google and #OpenAI
https://www.politico.eu/article/openai-google-reject-uks-ai-copyright-plan/
KI-Update: KI im Wissenschaftstest, ChatGPT löst Bilderboom aus, AMD Ryzen
Das "KI-Update" liefert werktäglich eine Zusammenfassung der wichtigsten KI-Entwicklungen.
LOL. OpenAI finally realized their AI voice assistant was unbearably annoying and interrupted CONSTANTLY.
Took them long enough.
But notice WHO gets the *real* upgrade ("improved personality"): PAYING users.
Free users just get slightly less interruption. Tells you everything about their priorities. Pay up or talk to the rude robot.
https://techcrunch.com/2025/03/24/openai-says-its-ai-voice-assistant-is-now-better-to-chat-with/
#OpenAI #ChatGPT #Freemium #Tech
So let me get this straight: Trump’s tariff formula (a) was the result of an AI query which (b) was misunderstood by a human and (c) misapplied to (d) incomplete data on the basis of (e) bad economic thinking to (f) crash the world economy
https://alecmuffett.com/article/113089
#ai #grok #llm #openai #tariffs #trump
What really happened with Sam Altman at OpenAI? Keach Hagey joins @leo Laporte, @jeffjarvis & @richcampbell to discuss her latest report—plus Amazon's Nova, xAI’s moves, and more AI headlines on Intelligent Machines 813.
#AInews #OpenAI
Download and subscribe here:
https://buff.ly/67HNwWE
heise+ | Privacy Policies generativer KI-Plattformen – ein Überblick
DeepSeek gilt als technisch bahnbrechend, doch wirft es Bedenken hinsichtlich des Datenschutzes auf. Unterscheidet sich DeepSeek darin von OpenAI oder Mistral?
OpenAI führt mehrere Klagen wegen seiner Praktiken bei Trainingsdaten und der Handhabung des Urheberrechts vor US-Gerichten. Daher wirft das Paper kein besonders schmeichelhaftes Licht auf das Unternehmen.
heise+ | Bildgenerator von ChatGPT 4o mit laxen Ethikrichtlinien
Der in ChatGPT 4o eingebaute Bildgenerator erzeugt glaubhaft gefälschte Porträts von Politikern und Animationsfilmen. Kann das gutgehen?
OpenAI unter Verdacht: GPT-4o angeblich mit O'Reilly-Büchern trainiert
OpenAI trainierte sein KI-Modell GPT-4o einer Studie zufolge mit Büchern des US-Technikverlages O'Reilly. Die Autoren fordern Lizenzverträge für KI-Trainings.
Le meilleur article sur #openai et Ghibli est sur @blast_info : https://www.blast-info.fr/articles/2025/images-generees-dans-le-style-de-ghibli-des-artistes-spolies-le-technofascisme-celebre-xtowO7zCTM67xyPOOUswiw
Sigh. You can tell when #mastodonau is being crawled by one of the #AI consumers by how long it takes to do stuff. When it's being left alone, everything is pretty much instant. When there's *three* crawlers attacking us at the same time, it takes 10-15 secs to do anything.
I promise I'm doing everything I can to stop them, but they don't listen to policy and authentication, so I have to manually blacklist things when I notice them. Feel free to @ me if mau is running slow, and I'll get the firewall filters out. #mastodon #screwyouai #stopmakingeverythingslow #openai
ChatGPT AI音声とおしゃべりしたら「まったくやる気が湧かないですね」「何か文句でも」との答えにどんよりした（CloseBox）
https://www.techno-edge.net/article/2025/04/03/4238.html
#technoedge #テクノロジー #ニュース #レビュー #ゲーム #ガジェット #OpenAI #ChatGPT #CloseBox
Researchers analyzed 13,962 paragraph excerpts from 34 #OReilly books, finding that #OpenAI #GPT4o "recognized" significantly more paywalled content than older models like GPT-3.5 Turbo. Technique, also known as a "membership inference attack," tests whether a model can reliably distinguish human-authored texts from paraphrased versions.
"GPT-4o [likely] recognizes, and so has prior knowledge of, many non-public O'Reilly books published prior to its training cutoff date"
https://techcrunch.com/2025/04/01/researchers-suggest-openai-trained-ai-models-on-paywalled-oreilly-books/
#OpenAI офіційно відкрили доступ до своєї #OpenAIAcademy для всіх користувачів.
Раніше компанія заявляла, що ініціатива була створена для підтримки розробників ШІ і бізнесу. Тепер освітні матеріали стали повністю безоплатними і доступні абсолютно всім охочим.
Основна частина програми присвячена навчанню навичкам промпт-інжинірингу та ефективному використанню продуктів OpenAI. Водночас для розробників додано курси з більш спеціалізованою тематикою.
Though Goro Miyazaki predicts that artificial intelligence could one day replace animators, he says it's impossible to replicate the depth of his father's work. https://www.japantimes.co.jp/culture/2025/04/02/film/ai-studio-ghibli-goro-miyazaki/?utm_medium=Social&utm_source=mastodon #culture #film #openai #chatgpt #ai #hayaomiyazaki #anime #studioghibli
Of course, #OpenAI did not pay for these books...
If you use ChatGPT, you're working with thieves.
OpenAIが400億ドルを調達、評価額は3,000億ドル
https://cafe-dc.com/hpc/openai-raises-40bn-valued-at-300bn/
#cafedc #aiデータセンター #altimeter #coatue #openai #stargate #thrive #ソフトバンク #マイクロソフト #生成ai
TechnoEdge 
