Daily digest

Today in AI

Tuesday, 18 August 2026

The top 8 of 11 AI stories in plain language, a prompt to try and one AI word to learn — about five minutes.

01
GPT-5 vs Gemini: Frontier AI vs Efficient Models
AI companies are increasingly balancing powerful frontier models with cheaper models for routine workloads, changing how businesses use AI.
Read the story →Source: cnbc.com
02
Claude & Agentic AI Security: New Risks for AI Agents
AI agents like Claude can use tools and take actions, creating new security risks around permissions, data access and autonomy.
Read the story →Source: genai.owasp.org
03
Sora & Creator Rights: Who Owns AI Training Data?
Creator content is becoming valuable AI training data, raising new questions about copyright, consent, licensing and creator compensation.
Read the story →
04
GPT-5: OpenAI Leadership Shake-Up
Open AI is reshuffling its leadership as it expands enterprise AI and prepares for its next major growth phase.
Read the story →Source: axios.com
05
Claude Workbench Retirement: What Developers Need to Know
Anthropic is retiring the legacy Claude Workbench and three experimental prompt APIs today, August 17, 2026.
Read the story →Source: platform.claude.com
06
Gemini & GPT-5: The AI Cost Optimization Race
Companies are rethinking AI spending as expensive frontier models drive higher inference costs, pushing businesses toward smarter model routing.
Read the story →
07
Claude vs Amazon AI: Who Owns Creator Content?
Twitch now lets creators opt out of Amazon generative-AI training, but the setting is enabled by default, sparking a major creator-rights debate.
Read the story →Source: twitch.tv
08
Claude AI & the Crisis of Trust in Artificial Intelligence
Anthropic CEO Dario Amodei says AI faces a major trust gap and argues that real-world benefits—not just impressive demos—will determine public confidence.
Read the story →