Welcome back to EmergingTech.
This week starts with a reality check for health AI.
A new study found that while leading AI models perform well on many medical tasks, they can still make confident mistakes when the input changes.
Let’s dive in.
LATEST IN AI
Anthropic Brings AI Agents Into Slack
Anthropic has launched Claude Tag, a beta feature that brings AI directly into Slack. Just tag @Claude to answer questions, summarize chats, write code, or help with tasks without leaving the conversation.
Powered by Claude Opus 4.8, Claude can follow conversations, connect to company tools, and work in the background. Since everyone can see its responses, teams can collaborate with the AI together.
The feature aims to make work faster and easier. But businesses will still need strong security controls to keep company data safe. Read more→
Agentic AI Is the Next Big Leap
AI is evolving beyond simple chatbots. The latest wave, known as Agentic AI, can plan, make decisions, use tools, and complete entire workflows with minimal human input.
The four stages of AI:
Machine Learning: Finds patterns and makes predictions.
Generative AI: Creates text, images, code, and more.
AI Agents: Completes multi step tasks using tools.
Agentic AI: Coordinates multiple tasks and manages end to end processes.
Why it matters: As AI moves from creating content to taking action, businesses could automate far more complex work than ever before.
AI IN HEALTH
Can You Trust AI With Your Health?
A new Nature Medicine study found that leading AI models like GPT 5 and Gemini perform well on many health tasks but still struggle when faced with small changes or missing information.
Researchers found the models could sometimes guess the right answer without key details, while minor prompt changes led to confident but incorrect responses. The findings suggest health AI is improving fast, but it is not yet ready to be relied on in every medical situation. Read more→
IEEE
AI Is Starting to Read Between the Lines
Source: IEEE
Emotion AI is becoming more aware of context instead of simply labeling people as "happy" or "sad." By analyzing facial expressions, voice, and body language together, these systems aim to better understand how someone is really feeling.
The technology is already being tested in areas like employee wellbeing and hiring. While it could help spot stress or burnout earlier, experts say it also raises important questions about privacy, accuracy, and how emotional data should be used. Read more→
IN RADAR
Amazon is investing another $13 billion in India, taking its total AI and cloud investment to $48 billion by 2030. The money will expand AWS data centers and AI infrastructure across the country. Read more→
India's first AICTE approved BTech in Artificial Intelligence and Data Science will combine modern AI skills with Indian Knowledge Systems. The programme also aims to support the digitisation of ancient manuscripts and preserve India's rich knowledge heritage. Read more→
OpenAI has limited the rollout of GPT 5.6 after a US government request. The company says the move is temporary and should not become the norm. Read more→
Boehringer Ingelheim is investing $15 million in Immunai to use AI to discover new drug targets for cancer and autoimmune diseases. Read more→
🌱 What are you building? Share with us
