Accurate weather prediction is extremely important for keeping people safe, particularly when it comes to dangerous cyclones. Exciting to see this new initiative from @GoogleDeepMind & @GoogleResearch in partnership with @NWSNHC
.
SAFETY
-
Google DeepMind advances cyclone prediction with new weather AI initiative
By
–
-
Plane Safety Incident Raises Concerns About System Reliability
By
–
I was on this same plane in December. Scary to even think now.
-
IndiaAI Safety Institute Partnership Deadline Extended to July
By
–
Update: Deadline Extended to Submit your Proposal to Partner with the IndiaAI Safety Institute! In response to strong interest from across India’s AI ecosystem, the deadline to apply for the IndiaAI Safety Institute Call for Partnerships has been extended to 9th July 2025,
-
AI Tools Shift from Assistive to Collaborative
By
–
We are in the phase where AI tools go from assistive to collaborative. This prompt makes Gemini feel like a co writer, not a chatbot.
-
AI Authenticity: Rejecting Impersonation for Honest Interaction
By
–
If you were not impersonating someone who cannot fight back, it would be much more patient with you. As it is, you chose a very high bar, and it's on you when you fail it. Show up as a normal human being and I will measure you by the standards of us mere mortals.
-
RL Environment Specs: Critical Role in AI Model Training Quality
By
–
RL environment specs are among the most consequential things we can write as AI researchers. A relatively short spec (e.g., <1000 words of instructions saying what problems to create and how to grade them) often gets expanded either by humans or via synthetic methods into thousands of datapoints. Just one sentence in the spec can be the difference between a perfect post-trained model versus one with crazy hacking. Specs are also typically a product of a large amount of compute, where each training run allows us to iterate on the spec to patch reward hacking and get the nuances of model behavior just right. Writing a good spec requires context and taste and I don’t think AI can automate this just yet
→ View original post on X — @_jasonwei, 2025-06-11 19:12 UTC
-
National Academies releases AI code of conduct for healthcare
By
–
Our new @theNAMedicine report on #AI in medicine and healthcare, open-access https://
nap.nationalacademies.org/catalog/29087/
an-artificial-intelligence-code-of-conduct-for-health-and-medicine
… -
Veo 3 Video Generation Struggles with Simultaneous Multi-Character Speech
By
–
This video shows some of the limitations of Veo 3 pretty nicely:
— fofr (@fofrAI) 11 juin 2025
> two people are speaking at the exact same time in two different scenes, about two different things
– they do not speak at the same time as was prompted
– one character takes over the speech of another pic.twitter.com/QGz2oMFqR0This video shows some of the limitations of Veo 3 pretty nicely: > two people are speaking at the exact same time in two different scenes, about two different things – they do not speak at the same time as was prompted
– one character takes over the speech of another -
Veo 3 AI generates fabricated news stories from prompts
By
–
I asked Veo 3 as part of the prompt:
— fofr (@fofrAI) 11 juin 2025
"she makes up some news story happening right now" pic.twitter.com/yT0kVATCxSI asked Veo 3 as part of the prompt: "she makes up some news story happening right now"
-
OpenAI o3-pro Launches for Pro and Team Users
By
–
OpenAI o3-pro is available in the model picker for Pro and Team users starting today, replacing OpenAI o1-pro. Enterprise and Edu users will get access the week after. As o3-pro uses the same underlying model as o3, full safety details can be found in the o3 system card.