What Happened
Diplomat-turned-Anthropic special envoy Jeff Bleich appeared at a fireside chat on artificial intelligence hosted by the Australian Strategic Policy Institute in Canberra on 14 September 2026. The event came barely a day after his boss, Anthropic CEO Dario Amodei, published an essay calling for "pacing" of AI development, including independent monitoring of AI models and global regulation. OpenAI bosses had made similar calls around the same time.
Bleich pushed back on any reading of "pacing" as a brake on development. "It's not as though pacing requires slowing down," he said. "If anything, you need to accelerate because you need even more compute for the research related to safety and staying ahead of the models."
Riffing on quotes from former US presidents, the former Washington ambassador to Australia said Anthropic welcomed doing business under robust regulation. He invoked Franklin D Roosevelt in a pointed aside: "The biggest thing we have to fear is fear itself." The reference was to Roosevelt's argument against the society-wide paralysis caused by massive economic upheaval, applied here to AI policy.
Bleich also addressed the forum's reception directly, drawing on another presidential register: "We didn't come here because it's easy. We came here because we knew you would make it hard."
Why It Matters
The remarks are notable because they came from inside Anthropic, the company whose CEO had just published the essay prompting concern about AI risks. Bleich's framing separates "pacing" from "slowing down" in a way that has direct implications for how Australian regulators and policymakers interpret calls for oversight. Australia's signals intelligence chief separately warned on 15 September 2026 that the country's outdated technology infrastructure is vulnerable to AI-enabled hacking attacks, according to The Guardian, adding a security dimension to the domestic policy debate.
Key Details
Australian advocates at the ASPI forum made the case for a specific national strategy. One speaker argued that Australia has historically succeeded by working the less fashionable parts of the value chain: "Australia has been very good as a nation at making the unfashionable end of the value chain work for us."
The same speaker pointed to concrete steps: "It means talking to the data centres that come here about bringing their researchers here; it means talking to the big AI companies that locate in Australia."
On the question of jobs, the forum heard a comparison to earlier technological transitions. "When cars came along, it was predictable that people wouldn't need as many horseshoes and stables," one speaker said. "On the other hand, there were all sorts of new jobs that came along with it."
Background and Context
Anthropic occupies an unusual position in the AI industry. The company was founded with an explicit safety focus and has consistently called for government oversight, yet it is also one of the most well-funded AI developers in the world. Bleich's role as special envoy involves engaging with governments, including Australia's, on policy questions. His appearance at ASPI placed him in front of an audience that includes defence, intelligence, and foreign policy professionals.
The ASPI forum sits within a broader Australian policy moment. The federal government has been consulting on AI regulation frameworks, and the Australian Signals Directorate has raised concerns about infrastructure vulnerability. The Guardian reported on 15 September 2026 that Australia's signals chief cited outdated technology as a specific risk factor for AI-assisted attacks.
What Comes Next
No specific legislative or regulatory announcements came out of the ASPI forum. Bleich's comments suggest Anthropic will continue to engage with Australian policymakers on the terms of any oversight regime. The signals chief's warning about infrastructure vulnerability, reported separately, points to a parallel track of government attention on defensive AI capabilities.