AI Safety Pacts and Agents

By the Orchestra editors · 4 min read

TL;DR

Australia and UK sign AI safety pact amid Anthropic Mythos concerns
CommBank launching AI agent to handle business loan applications
ClickUp replacing hundreds of staff with AI agents
Federal politician caught by WhatsApp phishing from foreign state actor

AI · Worldwide

ClickUp swaps hundreds of staff for AI agents

The project management startup laid off hundreds of employees and plans to replace them with thousands of AI agents. It's one of the starkest examples yet of a company explicitly substituting headcount with autonomous software.

Pope Leo's first encyclical targets unchecked AI power

Pope Leo XIV used his inaugural encyclical to call on governments to slow AI development and build new legal and ethical frameworks, warning of AI-driven warfare and labour displacement.

Australian Tech

Australia and UK sign AI safety pact

A new bilateral MOU will see Australia's forthcoming AI Safety Institute collaborate with its UK counterpart on frontier model risks, with Anthropic's Mythos model cited as a key concern.

CommBank building AI agent for business loans

Commonwealth Bank is rolling out an AI agent that handles income verification and other steps in business loan applications, a notable move by an AU big four bank into agentic workflows.

Federal politician compromised in WhatsApp phishing attack

A foreign state actor phished a federal MP and three staffers via WhatsApp in March, prompting parliament to block WhatsApp web access across its networks.