Australian Senators Call OpenAI and Anthropic Chiefs to Testify
Australia's Senate wants the chief executives of OpenAI and Anthropic to front an AI inquiry after a government website was hacked — a rare summons of foreign tech founders.

Australia's Senate has requested that the chief executives of OpenAI and Anthropic PBC appear before an artificial intelligence inquiry, following the recent revelation of a hack of an Australian government website.
Australia's Senate has asked the chief executives of OpenAI and Anthropic PBC to appear in person before a parliamentary inquiry into artificial intelligence, an escalation that follows the recent disclosure that a government website was hacked. The request was reported by Bloomberg Technology.
Both companies are privately held and neither trades on a public exchange, so there is no share price to register the shock. That does not make the moment small. Senate committees rarely ask the founders of foreign private companies to travel and sit for questioning, and the two names on the list are the two most visible builders of frontier AI models in the Western market. The request marks the point where an inquiry that had been about policy in the abstract attaches itself to a specific security incident.
Why a website breach pulled model developers into the room
The sequence matters. Legislative inquiries into artificial intelligence have generally circled the familiar themes — copyright, labour displacement, safety testing, children's exposure to chatbots. Those hearings tend to draw regional executives and policy staff rather than founders.
A hack of a government website is a different category of event. It is concrete, it is dated, and it has an owner somewhere in the chain. Once a security incident enters the record of an inquiry, the committee's questions stop being hypothetical: who had access, what tooling was involved, what was disclosed and when. That is the kind of questioning that lands on a chief executive rather than a government affairs manager, because the answers touch product decisions, not talking points.
Neither the identity of the affected site nor the precise mechanism of the breach is established in the reporting available so far, and it would be wrong to fill that in by inference. What is established is that Australian senators have connected the incident closely enough to frontier AI development to want the people who run OpenAI and Anthropic in front of them.
A request, not a subpoena — and the difference is the story
Parliamentary committees in Westminster-derived systems generally begin with an invitation. Compulsion exists, but it is awkward to apply to executives of foreign companies who are outside the jurisdiction for most of the year, and committees know it. So the first move is a request, issued publicly, and the pressure is reputational rather than legal.
That gives the companies a menu of responses, each of which reads differently:
- Attend in person. The strongest signal of cooperation, and the highest-risk option. Hearings are televised and hostile questions are the point.
- Appear by video link. The compromise that has become standard for overseas witnesses. It satisfies the form of the request while limiting exposure.
- Send a senior deputy. Common, and usually read by committees as a partial refusal.
- Submit written evidence only. The cheapest option in time and the most expensive in goodwill, and the one most likely to trigger talk of formal compulsion.
Which door the two firms walk through will tell regulators elsewhere how much leverage a mid-sized jurisdiction actually has over companies of this size.
Australia's regulatory track record makes this less of a long shot
Which door the two firms walk through will tell regulators elsewhere how much leverage a mid-sized jurisdiction actually has over companies of this size.
Australia is a small market by revenue and an outsized one by precedent. It has repeatedly written rules that larger jurisdictions later borrowed or at least studied closely — on the commercial relationship between platforms and news publishers, and on minimum ages for social media accounts. Global platforms that initially treated Canberra as a rounding error learned that the legislation travelled.
That history is the reason this request should not be filed under local politics. If a Senate inquiry produces findings that tie an AI developer's tooling to a breach of public infrastructure, those findings become source material for committees in other capitals. Evidence given once in Canberra is quotable everywhere.
What the public markets are — and are not — saying
Investors have no direct instrument for this. OpenAI and Anthropic are private, and the nearest public proxies are the infrastructure and platform companies that supply or distribute their models. Nothing in the most recent session suggests the AI complex is bracing for regulatory pain: the S&P 500, tracked by SPY, closed at $771.35, up 0.54% on the day from a prior close of $767.18, while the Nasdaq 100 proxy QQQ finished at $744.50, up 0.46% from $741.10. The Dow 30 proxy DIA closed at $517.49, a gain of 0.94%. Those are last traded prices as of 20:00 GMT on Friday, 25 September 2026, with the market closed.
The read-through is that a hearing request in a single foreign parliament does not move a benchmark, and should not be expected to. The transmission mechanism, if there is one, runs slower: through government procurement policy, through enterprise buyers who demand indemnities after a public-sector incident, and through the liability rules that legislators start drafting once a committee has a transcript in hand.
What to watch from here
Three things will determine whether this becomes a durable story or a news cycle. First, whether either chief executive agrees to appear, and in what format. Second, whether the inquiry's terms of reference are widened to cover the breach specifically, which would extend its timeline and invite evidence from Australian security agencies as well as from the vendors. Third, whether the committee's interim findings recommend anything binding — procurement restrictions, mandatory incident disclosure, or a statutory duty of care for model providers selling into government.
For the companies, the calculation is straightforward and uncomfortable. Declining to appear preserves a day in the diary and hands a committee a grievance it can use for months. Appearing means answering, under privilege and on camera, questions about a security incident whose full shape is not yet public.
Key facts
- Who was called: Chief executives of OpenAI and Anthropic PBC
- Requesting body: Australian Senate AI inquiry
- Trigger: Recent revelation of a hack of a government website
- Market backdrop: S&P 500 proxy SPY closed at $771.35, +0.54%, as of 20:00 GMT Sept 25, 2026
Frequently asked questions
What has the Australian Senate actually asked for?
The Senate has requested that the chief executives of OpenAI and Anthropic PBC appear before an inquiry into artificial intelligence to answer questions. The request follows the recent revelation of a hack of an Australian government website. It is a request for attendance rather than a confirmed hearing date, and the companies have a range of possible responses.
Are OpenAI or Anthropic publicly traded?
No. Both OpenAI and Anthropic PBC are privately held companies without a listing on a public stock exchange, so there is no share price that reacts directly to news such as a parliamentary summons. Investors seeking exposure to the sector generally hold the listed infrastructure, cloud and platform companies that supply or distribute frontier AI models.
Which government website was hacked?
The specific site has not been established in the reporting available so far. What is confirmed is that a hack of an Australian government website was recently revealed and that the Senate connected the incident closely enough to frontier artificial intelligence development to request testimony from the leaders of OpenAI and Anthropic.
Can the Senate force the executives to attend?
Parliamentary committees in Westminster-derived systems typically begin with an invitation rather than compulsion, and enforcing attendance against executives of foreign companies located outside the jurisdiction is legally and practically difficult. The immediate pressure is reputational. Common outcomes include appearing by video link, sending a senior deputy, or submitting written evidence instead.
Why does a hearing in Australia matter globally?
Australia has repeatedly written technology rules that larger jurisdictions later studied or adopted, including on platform payments to news publishers and minimum ages for social media. Testimony given to an Australian Senate committee enters the public record and can be cited by legislators elsewhere, so the practical reach of the inquiry extends well beyond the domestic market.
How did equity markets close before this news?
As of the last trade at 20:00 GMT on Friday, 25 September 2026, the S&P 500 proxy SPY closed at $771.35, up 0.54% from a prior close of $767.18. The Nasdaq 100 proxy QQQ closed at $744.50, up 0.46%, and the Dow 30 proxy DIA closed at $517.49, up 0.94%.
Sources
- Australia Senate Requests OpenAI, Anthropic CEOs Face AI Inquiry — Bloomberg Technology


