Yes, there is good news about AI, and late October 2026 had plenty of it: a new lab using AI to spot biological threats early, drug developers using AI to make better calls in cancer trials, and a useful research tool that stays free. These stories come from ClaudeGPT Si, a free and independent AI publication that links the sources behind every article, so you can check each one yourself.
Why the good stories are easy to miss
Headlines about AI tend to be about layoffs, lawsuits and leaks. Those stories matter, but they crowd out the quieter ones about AI doing something helpful. The four below were all reported in the same week, and each has a named source you can read.
1. A lab built to catch biological threats early
On October 24, the UK government launched the AI Biosecurity Lab, an independent non-profit. Its first job is early detection of emerging biological threats, the kind of work where a few days of warning can change how an outbreak goes.
The money is shared. The government puts in £100 million, and OpenAI and Anthropic together pledge up to £200 million, mostly as credits for their AI models. Google DeepMind is contributing scientists. It is a rare case of rival AI companies backing the same public project.
2. AI helping steer cancer drug trials
Pathos AI uses AI to guide cancer drug development: which patients a drug is most likely to help, when to change a trial, and which paths to drop early. On October 23 it announced a new funding round led by Bristol Myers Squibb, a large drugmaker that also gets access to Pathos's platform for its own neuroscience programs. ClaudeGPT Si's report on the round explains how the platform works.
When a drugmaker invests in a tool and starts using it the same day, it suggests the tool is doing real work, not just making for a nice press release.
3. Safer training at OpenAI
OpenAI said on October 23 that it now trains its most capable models with no live internet access. The models learn from a large store of cached web pages instead. It is a cautious step: a model being trained cannot reach out to real websites, so a mistake inside the training setup stays inside it.
4. A free research helper with a new name
Google renamed NotebookLM to Gemini Notebook in July 2026. Nothing was lost in the move: existing notebooks and shared links carry over. As of October 27, 2026, Google's plans page still lists Gemini Notebook on the free plan, according to ClaudeGPT Si's explainer. Students and anyone who reads a lot of documents can keep using it at no cost.
Good news at a glance
| Story | Who | What it does |
|---|---|---|
| AI Biosecurity Lab | UK government, OpenAI, Anthropic, Google DeepMind | Early detection of biological threats |
| Pathos AI | Pathos, Bristol Myers Squibb | AI guidance for cancer drug trials |
| Offline training | OpenAI | Keeps models in training away from the live web |
| Gemini Notebook | Free research notes tool, renamed from NotebookLM |
How to keep finding stories like these
The simplest way is to read a source that reports both kinds of news from the original announcements. ClaudeGPT Si is free to read, shows when each article was last updated, and keeps an RSS feed you can follow. Its sections split AI news from apps, research and guides, so you can head straight for the hopeful science stories if that is your mood.
It will not filter out the bad news, and it should not. A fair picture needs both. If you only want cheerful headlines, a dedicated good-news site is the better fit.
And when you have read enough about AI for one day, Mayavoo has free browser games with nothing to install.