The Weekly AI Digest
Week of 27 February – 5 March 2026
In AI News This Week
Anthropic Blacklisted by Pentagon, Then Hits #1 in App Store
The Pentagon designated Anthropic a “supply chain risk” after the company refused to remove safety guardrails. President Trump ordered all federal agencies to phase out Anthropic’s technology within six months. Hours later, Claude became the #1 downloaded app on Apple’s App Store, overtaking ChatGPT.
Why it matters: This is the first major test of what “safety-first AI” means when government contracts are on the line. The split between OpenAI and Anthropic now represents a clear philosophical divide.
Block Cuts 4,000 Jobs, Says AI Makes Smaller Teams Possible
Jack Dorsey’s Block laid off nearly 40% of its workforce, explicitly tying the cuts to AI. Block’s stock surged 17%. Even Sam Altman acknowledged: “There’s some AI washing where people are blaming AI for layoffs that they would otherwise do.”
Why it matters: Whether or not AI drove these specific cuts, Dorsey’s framing signals how executives will justify workforce reductions going forward.
OpenAI Releases GPT-5.3 Instant to Fix “Cringe” Problem
OpenAI released GPT-5.3 Instant, focused on making ChatGPT less “preachy.” The update claims a 27% reduction in hallucination rates and significantly reduces unnecessary refusals.
Why it matters: This reflects a shift across the industry from raw capability to user experience. The models are getting smarter, but they also need to stop being annoying.
Andrew Ng Says AGI Is Decades Away
AI pioneer Andrew Ng pushed back on AGI hype, saying it remains “many decades away.” PwC’s 2026 CEO Survey found that 56% of 4,454 CEOs reported neither increased revenue nor reduced costs from AI in the past year.
Why it matters: Ng is one of the most credible voices in AI. His scepticism about near-term AGI is a useful counterweight to vendor marketing.
London Hosts Largest Anti-AI Protest Yet
A few hundred protesters marched through London’s King’s Cross tech hub. The march was organised by activist groups Pause AI and Pull the Plug, covering concerns from “online slop” to autonomous weapons.
Why it matters: Public scepticism of AI is growing, but it hasn’t yet coalesced into a coherent political force.
Australia Watch
AI Age Verification Rules Take Effect March 9
Australia’s eSafety regulator will require AI platforms to prevent under-18s from accessing harmful content, with fines up to $49.5 million. Over half of the 50 most popular AI tools had taken no visible steps to comply.
Large platforms including ChatGPT and Claude have begun rolling out age assurance systems.
AI Safety Institute Debate Continues
Australia’s new AI Safety Institute is facing competing visions. Three groups are jockeying to shape its direction: near-term harms, existential risk and international coordination.
The institute currently lacks enforcement powers, operating more as an advisory body.
Healthcare AI Stuck in Pilot Phase
Only 12% of Australian healthcare organisations have fully deployed AI across multiple functions, while 60% remain in pilot stage.
The gap between experimenting with AI and getting real value from it remains wide.
Emerging Trends
The AI-Jobs Narrative Is Splitting in Two
Block’s layoffs and the Pentagon standoff represent two very different AI stories. One is about efficiency and displacement. The other is about values and control.
Companies need to decide not just what AI can do for them but what they’re willing to accept in exchange.
Tone and Trust Are Becoming Competitive Factors
OpenAI’s focus on making ChatGPT less “cringe” reflects a maturing market. Users want tools that feel natural to use.
For businesses building AI-powered products, user experience matters as much as technical benchmarks.
Expect More AI Safety Theatre
Australia’s age verification deadline will likely expose how few AI companies have built compliance into their products.
Businesses relying on third-party AI tools should check whether their providers are prepared for regulatory requirements.