BTL Topic index
A living archive
AI Governance
Policy, ethics, safety, and the human questions around AI deployment.
The working lens
Look for the system beneath the story.
Ideas earn their place here by being useful after the presentation ends and the real work begins.
01 Follow the consequences
02 Keep people in the system
03 Test it against reality
The working index
Ideas worth carrying into the work.
Start with the lead essay. Then follow whichever question feels most useful now.
-
Decoding Animal Communication: A Better Way to See Hidden Risk
Open article ↗: Decoding Animal Communication: A Better Way to See Hidden RiskMachine learning is being pointed at whale song and ape gesture with real money behind it. A new Topoi paper by Marriah Alcantara and Kristin Andrews sets out…
-
Reliable AI Systems: A Practical Way to Control Risky Outputs
Open article ↗: Reliable AI Systems: A Practical Way to Control Risky OutputsReliable AI systems are built around uncertain models. Here is the engineering discipline needed for context, agents, evaluation and production.
-
AI Job Design: How to Make Agents Earn Their Place
Open article ↗: AI Job Design: How to Make Agents Earn Their PlaceAI job design is the step most enterprises skip: no role, no scope, no decision rights, no named owner. What an agent needs before it starts work.
-
AI Leadership Judgement: How to Make Better Decisions With AI
Open article ↗: AI Leadership Judgement: How to Make Better Decisions With AIAI leadership judgement needs more than human approval. Use a practical review method and free pack to preserve independent decisions with AI.
-
AI Build Versus Buy: How to Make Better Ownership Decisions
Open article ↗: AI Build Versus Buy: How to Make Better Ownership DecisionsAI build versus buy is now an ownership decision. Compare whole-life obligations and use the free decision pack before approving agent-built software.
-
AI Is Not a University. But It Can Build You One.
Open article ↗: AI Is Not a University. But It Can Build You One.AI has made knowledge abundant, but learning still needs judgement. The ALTER framework turns AI into five practical roles, with a downloadable six-week course builder.
-
The Hidden Verification Gap: A Practical 30-Day Way Forward
Open article ↗: The Hidden Verification Gap: A Practical 30-Day Way ForwardThe verification gap appears when AI output outruns review. Use this practical 30-day plan to strengthen governance, quality and team learning.
-
Generative AI Customer Insights: What AI Can and Can’t Do
Open article ↗: Generative AI Customer Insights: What AI Can and Can’t DoGenAI can help organisations find and synthesise existing customer research. It cannot repair weak evidence or replace the analysis that creates trustworthy insight.
-
Seeing the Full Story Isn’t a Data Problem. It’s an Empathy Problem.
Open article ↗: Seeing the Full Story Isn’t a Data Problem. It’s an Empathy Problem.More data rarely tells you who you’re not reaching. Empathy and humility do — and they’re the precondition for the ladder of inference, not an alternative to it.
-
Token Spend Is Not a Cost to Cut. It Is a Portfolio to Manage.
Open article ↗: Token Spend Is Not a Cost to Cut. It Is a Portfolio to Manage.Cutting the AI bill misses the point. The organisations getting value from AI treat token spend as a portfolio to allocate, diversify and rebalance.
Change the lens
The useful questions refuse to stay in their lane.
Move sideways through the archive. The strongest connections often appear between categories.
Still carrying a question?