By Darian Macey, June 30 2025
Australia’s defence leaders have drawn ire over the past year for their infuriating habit of giving vague, non-answers – so-called “nothing statements.” These statements often dodge the question, fail to inform, and leave the public more confused. As today is Defence’s annual AI Summit, it’s worth examining five recent examples of this trend, across a range of defence issues, and imagining how those responses could have been much better if guided by AI-informed advice. Each case illustrates how AI tools could help craft clearer, more informative statements that actually address the questions at hand.
1. Marles’s VIP Flights and Golf Clubs Non-Answer
In September 2023, Deputy Prime Minister and Defence Minister Richard Marles was grilled in Parliament over his use of RAAF VIP flights – in particular, whether he ever packed his golf clubs on official trips. Instead of a straightforward answer, Marles offered a deflection. He failed to respond to the specific question, insisting only that every flight he’s taken was in pursuit of his duties and “absolutely” within the rules. This generic retort neither confirmed nor denied the golf-club query and left the issue dangling. Little wonder the exchange sparked criticism; it exemplified the “nothing statement” – technically saying words, but shedding no new light.
How AI could have helped: An AI-powered advisor could have analysed public sentiment and similar past controversies to warn that dodging the question would only fuel suspicions. AI could suggest a more direct, confidence-building reply. For instance, Marles could have stated clearly “I have not used Defence flights inappropriately – I’ve only carried equipment necessary for official duties.” He might even pre-emptively offer to release flight logs (a detail an AI system could quickly retrieve) to back up his claim. Such a fact-based, transparent answer would address the question head on. Rather than a lawyerly evasion, an AI-informed statement could reassure the public that nothing untoward was happening on those VIP flights, thus quelling the uproar before it started.
2. Stonewalling on AUKUS Submarine Refunds
The multi-billion-dollar AUKUS submarine program has come under intense scrutiny, yet Defence leaders have often been less than forthcoming. A glaring example came in June 2024, when Vice Admiral Jonathan Mead – head of the Australian Submarine Agency – was asked in Senate Estimates whether Australia’s $4.7 billion payment to the U.S. for AUKUS submarines had a refund clause if no submarines are delivered. Instead of a yes-or-no answer, Mead repeatedly refused to say if a refund safeguard existed. He fell back on a talking point: “the US has committed to providing two nuclear-powered submarines” – essentially a non-answer. Senators pressed, framing it as a hypothetical scenario, but Mead would not “entertain” the question. This obstinate stonewalling did nothing to assure Australians that their billions wouldn’t vanish if the deal fell through.
How AI could have helped: AI could assist advisors by simulating this Q&A and flagging that simply reciting the commitment line would come off as evasive. A more satisfying, AI-crafted response might have acknowledged the concern: “It’s highly unlikely the commitment will fail; however, if it did, we have arrangements to protect Australia’s investment.” Even if certain details were classified, AI could help formulate wording that signals contingency plans exist – without breaching sensitivities. By learning from similar defence contracts globally, an AI system might suggest referencing precedent (e.g. other deals that include protections) to strengthen the statement. The result? A response that actually addresses the refund question, reassuring the public that Defence isn’t writing a blank check, instead of the opaque nothing statement that raised more questions than it answered.
3. Concealing the True Cost of AUKUS (USD vs AUD)
Nothing statements aren’t always just verbal dodges – sometimes they take the form of strategic omission. A case in point is how Defence leaders communicated the cost of Australia’s AUKUS “buy-in.” For months after the AUKUS submarine plan was announced, ministers and officials cited “$3 billion” as Australia’s contribution – carefully not mentioning that this figure was in US dollars, not Australian. This omission misled the public, since US$3 billion is about AU$4.7 billion – over 50% more than the impression given. It wasn’t until a late-night Senate hearing in October 2023 that a Defence official finally admitted on record that the $3 billion was USD. The fact that it took so long for this currency detail to be clarified led to accusations of deliberate obfuscation. As Asia-Pacific Defence Reporter scathingly put it, officials “hidden the fact” by omitting the currency and “never corrected or clarified [it]” for at least eight months. This eroded trust – a classic own-goal for transparency.
How AI could have helped: An AI-driven communications aide would have caught this currency sleight-of-hand immediately. AI tools excel at spotting inconsistencies and potential public confusion. If advisors had run the AUKUS announcement through an AI clarity check, it would have flagged: “Mention currency to avoid misunderstanding.” With that prompt, Defence leaders could have proactively stated, “Australia’s contribution is US$3 billion – roughly AU$4.7 billion at the current exchange rate.” Such a statement, while perhaps a tough pill politically, would have earned credibility. AI could also draw on financial data to show the impact of exchange rates, equipping the minister to explain why quoting USD made sense (if it did) or at least to be candid about the true cost. In short, AI would urge transparency over spin. By front-footing the full AUD figure, leaders could have pre-empted the “deceit” narrative and treated Australians with respect – something far more satisfying than the misleading nothing statements we got.
4. Mixed Messages on Chinese Live-Fire Drills
When Chinese warships conducted surprise live-fire exercises off Australia’s coast in early 2025, the incident called for clear, calm communication. Instead, the government’s messaging turned into a confusing tangle. Prime Minister Anthony Albanese initially suggested that Australian authorities were on top of it, saying Australia received civilian and military notifications “at around the same time.” Yet the very next day, the ADF chief told Senate Estimates that the military only learned of the drill nearly 40 minutes after a commercial airline pilot’s warning. The timelines didn’t match up. Opposition leader Peter Dutton seized on this discrepancy, accusing Albanese of “ducking and weaving” and failing to answer basic questions about who knew what, when. The public was left unsure whether Defence was caught flat-footed or not, all because of a muddled, possibly evasive initial statement from the top.
How AI could have helped: In a fast-moving situation like this, AI systems could aggregate real-time data from Defence, Air Traffic control, and other sources to build a precise timeline for leaders before they speak. If Albanese’s advisors had an AI-generated brief, it would have highlighted the gap between the pilot’s alert and Defence’s awareness – warning that any attempt to gloss over it would backfire. AI could have recommended a frank, factual approach: for example, “We learned of the Chinese exercise with a short delay – a commercial pilot alerted us first, and Defence was notified shortly after. While this notice was later than ideal, we immediately took action and have raised our concerns with Beijing.” Such a statement would admit the delay (which was going to come out anyway) and focus on the response, rather than create a false impression of simultaneity. By aligning the public message with verified data, AI would help avoid the credibility gap that comes from “nothing statements.” In this case, a bit of algorithmic help could have kept the PM’s narrative consistent with the ADF’s, avoiding public confusion in the heat of a security incident.
5. The KPMG Data Project Cover-Up
Defence’s penchant for opaque statements even extends to internal management issues. Consider the controversy over a $100 million “One Defence Data” project contracted to KPMG. An independent review found the program was riddled with failures – including a finding that Defence paid KPMG for work that hadn’t been delivered. Yet when Senator David Pocock probed this in early 2024, Defence flatly denied anything was amiss. Departmental official Matt Yannopoulos testified that “no milestone payment was paid without the services being delivered to a suitable quality,” rejecting the review’s claim. This confident dismissal turned out to be misleading. Pocock later obtained internal emails showing Defence knew the work was subpar but decided to pay KPMG anyway “to maintain a collaborative relationship,” effectively validating the review’s findings. In other words, Defence’s initial statement to the Senate was a “nothing statement” at best – and at worst, verged on a cover-up. It certainly did not address the core issue of accountability, and only further eroded trust when the truth emerged.
How AI could have helped: An AI system scanning Defence’s own correspondence and the review report would have raised red flags about contradicting documented facts. Advisors armed with AI might have urged a different approach: instead of denial, acknowledge and contextualise. For instance, an AI-suggested statement could be: “In one case we approved a payment despite incomplete deliverables – a decision we made to preserve the schedule, but in hindsight it was not ideal and we’ve tightened oversight to ensure it doesn’t recur.” Such a candid admission, backed by the data an AI had surfaced, would address the question honestly. It’s a harder path, but ultimately less damaging than the obfuscation. By cross-referencing the review and internal emails, AI could have helped Defence come clean upfront. The improved statement would show accountability and a plan to fix the issue – leaving the public and Senate far more satisfied than a dismissive denial that falls apart later. In short, AI could inject some much-needed integrity into Defence communications, replacing the instinct to “say nothing to see if it blows over” with an ethos of “face the facts and fix them.”
Conclusion: Australia’s defence leadership can no longer afford “nothing statements” that insult the public’s intelligence. As we’ve seen, each vague, non-committal answer only generates frustration, whether the topic is ministerial travel, multibillion-dollar projects, foreign military incidents, or internal failures. The promise of AI at this Defence AI Summit isn’t just about cutting-edge tech on the battlefield – it’s also about better informing our leaders off the battlefield. From natural language processing that ensures clarity, to data-mining tools that surface the truth, AI could be the secret weapon to transform empty platitudes into meaningful communication. Had the above five statements been crafted with a bit of AI-assisted foresight, the Defence Department might have earned praise for transparency and substance instead of criticism for evasiveness. In an era where trust is as crucial a defence asset as any submarine or missile, embracing AI to sharpen public accountability may be one of the smartest moves our Defence leaders can make.
Sources: Recent Senate Estimates transcripts, media reports and expert commentary have highlighted these issues – including coverage by The Guardian, ABC News, and Asia-Pacific Defence Reporter, among others, which provided the details and critiques cited above.
Author: Darian Macey
Darian is a former Engineering Officer with the RAE with tertiary qualifications in Engineering and Project Management. He is a leading practitioner within Australia's NSI sector with previous employment in Government, project management and building contractor organisations.
