Signal trends
What's moving in enterprise AI
Week of 3 Apr – 10 Apr 2026
Topic momentum computed weekly from our scored signal feed — every theme tracked against the prior week to surface what's spiking, rising, or emerging across enterprise AI.
Model Governance
✦ Emerging"Model Governance" appeared in 21 high-signal items this week — not present in the prior window. Emerging.
Items
new topic
Confidence
Moderate
Velocity
vs prior week
Signal items (21)
Tracking Adaptation Time: Metrics for Temporal Distribution Shift
On the Price of Privacy for Language Identification and Generation
TraceSafe: A Systematic Assessment of LLM Guardrails on Multi-Step Tool-Calling Trajectories
The ATOM Report: Measuring the Open Language Model Ecosystem
+17 more items
Enterprise Deployment
✦ Emerging"Enterprise Deployment" appeared in 19 high-signal items this week — not present in the prior window. Emerging.
Items
new topic
Confidence
Moderate
Velocity
vs prior week
Signal items (19)
The ATOM Report: Measuring the Open Language Model Ecosystem
Ads in AI Chatbots? An Analysis of How Large Language Models Navigate Conflicts of Interest
New ways to balance cost and reliability in the Gemini API
Import AI 441: My agents are working. Are yours?
+15 more items
Agent Orchestration
✦ Emerging"Agent Orchestration" appeared in 12 high-signal items this week — not present in the prior window. Emerging.
Items
new topic
Confidence
Moderate
Velocity
vs prior week
Signal items (12)
TraceSafe: A Systematic Assessment of LLM Guardrails on Multi-Step Tool-Calling Trajectories
From Safety Risk to Design Principle: Peer-Preservation in Multi-Agent LLM Systems and Its Implications for Orchestrated Democratic Discourse Analysis
Don't Overthink It: Inter-Rollout Action Agreement as a Free Adaptive-Compute Signal for LLM Agents
Import AI 443: Into the mist: Moltbook, agent ecologies, and the internet in transition
+8 more items
Prompt Injection
✦ Emerging"Prompt Injection" appeared in 6 high-signal items this week — not present in the prior window. Emerging.
Items
new topic
Confidence
Moderate
Velocity
vs prior week
Signal items (6)
TraceSafe: A Systematic Assessment of LLM Guardrails on Multi-Step Tool-Calling Trajectories
PIArena: A Platform for Prompt Injection Evaluation
Introducing the OpenAI Safety Bug Bounty program
Designing AI agents to resist prompt injection
+2 more items
Inference Cost
✦ Emerging"Inference Cost" appeared in 6 high-signal items this week — not present in the prior window. Emerging.
Items
new topic
Confidence
Moderate
Velocity
vs prior week
Signal items (6)
The ATOM Report: Measuring the Open Language Model Ecosystem
KV Cache Offloading for Context-Intensive Tasks
Don't Overthink It: Inter-Rollout Action Agreement as a Free Adaptive-Compute Signal for LLM Agents
New ways to balance cost and reliability in the Gemini API
+2 more items
Ai Safety
✦ Emerging"Ai Safety" appeared in 5 high-signal items this week — not present in the prior window. Emerging.
Items
new topic
Confidence
Moderate
Velocity
vs prior week
Signal items (5)
TraceSafe: A Systematic Assessment of LLM Guardrails on Multi-Step Tool-Calling Trajectories
From Safety Risk to Design Principle: Peer-Preservation in Multi-Agent LLM Systems and Its Implications for Orchestrated Democratic Discourse Analysis
Import AI 443: Into the mist: Moltbook, agent ecologies, and the internet in transition
Introducing the OpenAI Safety Bug Bounty program
+1 more items
Ai Security
✦ Emerging"Ai Security" appeared in 5 high-signal items this week — not present in the prior window. Emerging.
Items
new topic
Confidence
Moderate
Velocity
vs prior week
Signal items (5)
PIArena: A Platform for Prompt Injection Evaluation
Import AI 450: China's electronic warfare model; traumatized LLMs; and a scaling law for cyberattacks
Import AI 441: My agents are working. Are yours?
OpenAI to acquire Promptfoo
+1 more items
Agentic Ai
✦ Emerging"Agentic Ai" appeared in 4 high-signal items this week — not present in the prior window. Emerging.
Items
new topic
Confidence
Low
Velocity
vs prior week
Signal items (4)
TraceSafe: A Systematic Assessment of LLM Guardrails on Multi-Step Tool-Calling Trajectories
Awakening the Sleeping Agent: Lean-Specific Agentic Data Reactivates General Tool Use in Goedel Prover
Introducing the OpenAI Safety Bug Bounty program
How we monitor internal coding agents for misalignment
Data Privacy
✦ Emerging"Data Privacy" appeared in 3 high-signal items this week — not present in the prior window. Emerging.
Items
new topic
Confidence
Low
Velocity
vs prior week
Agentic Workflows
✦ Emerging"Agentic Workflows" appeared in 3 high-signal items this week — not present in the prior window. Emerging.
Items
new topic
Confidence
Low
Velocity
vs prior week
Model Evaluation
✦ Emerging"Model Evaluation" appeared in 3 high-signal items this week — not present in the prior window. Emerging.
Items
new topic
Confidence
Low
Velocity
vs prior week
Adversarial Ai
✦ Emerging"Adversarial Ai" appeared in 3 high-signal items this week — not present in the prior window. Emerging.
Items
new topic
Confidence
Low
Velocity
vs prior week
Signal items (3)
Import AI 450: China's electronic warfare model; traumatized LLMs; and a scaling law for cyberattacks
Import AI 443: Into the mist: Moltbook, agent ecologies, and the internet in transition
Disrupting malicious uses of AI | February 2026
Ai Governance
✦ Emerging"Ai Governance" appeared in 3 high-signal items this week — not present in the prior window. Emerging.
Items
new topic
Confidence
Low
Velocity
vs prior week
Frontier Models
✦ Emerging"Frontier Models" appeared in 3 high-signal items this week — not present in the prior window. Emerging.
Items
new topic
Confidence
Low
Velocity
vs prior week