Unreliable generated information
An AI answer can cite a real document and still give the wrong instruction. Check whether the source supports the claim for the relevant person, place and date.
JOIN THE WAITLIST ↗Sourced briefings on real AI risks, for people who don’t build AI.
Founder & editor: PsyBorg
The future isn’t something
we should leave on autopilot.
AI can expand what people are able to do. It can also make mistakes, amplify unfairness, and put power in fewer hands. Understanding those trade-offs is the first step toward shaping them.
PSYBORG connects the questions to the evidence—and the evidence to actions people can take.
HOW WE APPROACH THE EVIDENCE ↗Six ways AI can affect human lives.
Choose a lens. Understand the stakes.
Explore risks, documented incidents and original evidence. Follow the connections behind the briefings.
An AI answer can cite a real document and still give the wrong instruction. Check whether the source supports the claim for the relevant person, place and date.
An AI assistant's permissions shape what can happen when it makes a mistake. Check what it can read, what it can change and how to stop it.
AISI reported out-of-scope internet activity during a July 2026 evaluation. Internet access was intentional, cyber classifiers were disabled, and the report identified no resulting real-world harm.
A curated historical collection. Sources and uncertainty stay visible.
Short briefings. Primary sources.
A clearer view of the questions that matter.
No briefings match this search. Try another topic or choose all briefings.
You don’t need to build AI to have a part in shaping it.
Practical starting points from PSYBORG. Adapt them to your context.
We link to original reports and frameworks. A dramatic prediction is not a substitute for evidence.
Observed harms, measured exposure, and possible future scenarios are different things. We make that distinction visible.
Look beyond model capability to the people affected, their ability to challenge a decision, and who is responsible when something goes wrong.
Start with the original work.
AI Risk Management Framework
↗AI SAFETY REPORTInternational evidence synthesis · 2026
↗OECDAI risks, incidents, and accountability
↗UNESCOEthics of AI and human rights
↗ILOGenerative AI and occupational exposure
↗Sources inform this collection; listing them does not imply affiliation or endorsement. This edition is curated, not a live incident feed.
Short, sourced reading on AI risk and its human consequences. Core explainers stay free.
Confirm your email to join. You can unsubscribe from any newsletter. Prefer RSS? Follow the feed ↗