How Artificial Intelligence works and plans to phase out animal testing

OpenAI has built an experimental large language model that is easier to interpret, while the UK has announced timelines to end many forms of animal testing backed by new nonanimal technologies.

OpenAI has developed an experimental large language model that is notably easier to understand than typical models. The story frames this as a significant step because most large language models remain black boxes, limiting researchers’ ability to explain why models hallucinate, veer off course, or should be trusted with high-stakes tasks. The piece positions transparency as a route to better diagnostics and safer deployment of Artificial Intelligence in critical applications.

Google DeepMind is advancing in a complementary area by combining language models with embodied agents. The company built SIMA 2, a video-game-playing agent that operates in 3D virtual worlds such as Goat Simulator 3. SIMA 2 is built on top of Gemini, DeepMind’s flagship large language model, which the company says gives the agent a substantial boost in capability. The work is presented as progress toward more general-purpose agents and improved real-world robotics, building on an earlier demo of SIMA that DeepMind showed last year.

The newsletter also highlights a major policy shift in the United Kingdom, where the science minister announced an ambitious plan to phase out animal testing. Specific milestones include ending tests for potential skin irritants by the end of next year, expecting to end tests of Botox strength on mice by 2027, and reducing drug tests in dogs and nonhuman primates by 2030. The announcement is framed as timely because recent advances in technologies that model the human body offer alternative ways to test potential therapies without animals. The change is described as welcome news for both activists and researchers opposed to animal experimentation.

70

Impact Score

Nvidia to sell fully integrated Artificial Intelligence servers

A report picked up on Tom’s Hardware and discussed on Hacker News says Nvidia is preparing to sell fully built rack and tray assemblies that include Vera CPUs, Rubin GPUs and integrated cooling, moving beyond supplying only GPUs and components for Artificial Intelligence workloads.

Navigating new age verification laws for game developers

Governments in the UK, European Union, the United States of America and elsewhere are imposing stricter age verification rules that affect game content, social features and personalization systems. Developers must adopt proportionate age-assurance measures such as ID checks, credit card verification or Artificial Intelligence age estimation to avoid fines, bans and reputational harm.

Large language models require a new form of oversight: capability-based monitoring

The paper proposes capability-based monitoring for large language models in healthcare, organizing oversight around shared capabilities such as summarization, reasoning, translation, and safety guardrails. The authors argue this approach is more scalable than task-based monitoring inherited from traditional machine learning and can reveal systemic weaknesses and emergent behaviors across tasks.

Contact Us

Got questions? Use the form to contact us.

Contact Form

Clicking next sends a verification code to your email. After verifying, you can enter your message.