A conservative lobby group's effort to defeat Victorian Labor has exposed an uncomfortable disagreement on the right, with some Coalition figures warning that the campaign could undermine their own bid for government. The group, Advance, is promoting a put-Labor-last message using advertising that includes AI-generated material. Supporters see a way to shift voters away from Labor; critics inside the Coalition worry the message may divide the anti-Labor vote rather than strengthen the Liberal and National parties.
The concerns were reported by the Guardian, which spoke to Coalition figures about the campaign. It is important to distinguish those reported views from an official declaration by the entire Coalition. The critics described in the report do not establish that every Liberal or National candidate opposes Advance. Nor is the story evidence of a formal split in the parliamentary party. It is a dispute about campaign strategy and the possible consequences of political advertising.
Advance plans to spend about $350,000 on a campaign targeting 20 marginal Labor-held seats, according to the report. The strategy combines digital and traditional advertising and uses messaging that encourages voters to rank Labor last. The geographic target matters. Marginal seats can be highly sensitive to small shifts in voter behaviour, and campaigns in those areas may be judged by whether they help or complicate a particular candidate's chance of winning.
The Coalition's central pitch is different from Advance's. Liberal and National campaigners want voters dissatisfied with the government to support Coalition candidates as the alternative. A message focused only on opposing Labor can leave unanswered which other party should receive a first preference. That is the source of the tension described by those expressing concern. An advertisement can be fiercely anti-government without being an advertisement in favour of the Coalition.
Some Coalition voices fear that Advance's approach could advantage One Nation, which competes for voters on the conservative side of politics. That is a political assessment reported by the Guardian, not a demonstrated election result. No ballot has been counted in the forthcoming election, and current evidence cannot establish that the campaign has cost the Coalition seats. The point is more limited but still significant: groups with a shared desire to remove Labor may disagree sharply over the most effective path to achieve it.
Advance has pushed back against the criticism. It argues its approach is grounded in electoral strategy and data about overlap between groups of voters. On that view, a broad effort to persuade people to put Labor behind other options can help produce a change of government even if a person does not start by voting Coalition. The group does not accept that its advertising is necessarily harming the opposition. Its response is essential to the story because the report concerns the anticipated effect of a campaign, not a settled finding of failure.
Australia's preferential voting system makes the debate more complicated than a simple contest over first preferences. Voters place candidates in order, and preferences can flow from eliminated candidates according to ballot choices. A first-preference vote for a minor party does not always mean the major conservative candidate loses that support entirely. Equally, the existence of preferences does not make the initial party choice irrelevant. Campaigns must consider which candidates survive each count and how voter choices are distributed in individual electorates.
The mechanics help explain why rival strategists can look at the same anti-Labor sentiment and reach different conclusions. A lobby group seeking a change in government might see value in persuading people to rank Labor last, regardless of their first choice. A Coalition candidate fighting a close seat might prefer a clear first-preference appeal, hoping to avoid confusion and to maximise the party's position through the count. Neither argument can be declared correct in advance without seat-specific evidence.
The use of AI-generated advertisements adds a second debate. Modern campaigns can produce provocative visual material cheaply and distribute it quickly. That does not automatically make an advertisement false or unlawful. But when artificial imagery is used in politics, viewers may struggle to distinguish an illustration from a record of something that actually occurred. Campaigners have a responsibility to ensure fabricated scenes and synthetic representations do not become false evidence about an opponent's actions.
The Guardian's report describes AI material in the campaign but does not justify inventing any particular image or dialogue that it did not document. It would be misleading to depict a real Victorian politician saying a line manufactured by a generator and imply the person made that statement. The existence of political AI advertising is newsworthy in its own right; the ethical question can be explained without spreading an altered advertisement as though it were authentic footage.
The controversy also reflects the changing relationship between political parties and outside advocacy organisations. Third-party campaigns can reach people who do not follow a party's official communications, choose a different tone and take risks that candidates might avoid. They can also make strategic choices a party cannot fully control. Sharing a broad political objective does not mean agreeing on messaging, candidate priority or the handling of rival parties drawing votes from a similar pool.
That independence can be useful for political participation, but it creates headaches when a party wants a disciplined election message. Coalition candidates may be asked about ads they did not commission and a strategy they did not design. Their responses can then become part of the public debate, potentially diverting attention from local policy issues. None of that establishes that Advance is acting on the Coalition's instructions. The reported disagreement is precisely about the gap between what the lobby group is doing and what some Coalition figures want.
Marginal electorates also vary. A voter attracted to a minor party in one district may behave differently in another, depending on candidates, local issues and the composition of the ballot. That variation makes statewide claims about a campaign's effect difficult to prove. A spending figure and a list of targeted seats tell us what an organisation plans to do, not how voters will respond. Any claim that the campaign will deliver a majority or inevitably cost one would go beyond the available evidence.
The Victorian election is taking place against wider concern over living costs, debt, transport and political trust. Those issues give parties plenty of reasons to compete for attention, and a sharp attack ad may be attractive in an information environment crowded with messages. Yet voters also look for practical answers about government. If an advertisement says who should come last but not who should govern, candidates may worry it leaves an important part of their argument unfinished. That is an interpretation of the strategic disagreement, not a claim about the private motives of every person involved.
The warning from some Coalition figures is therefore a question to the anti-Labor campaigners: is weakening the government enough, or must the effort also build support for a specific alternative? Advance's answer is that shifting preferences away from Labor serves the broader objective. Those positions differ even if both groups say they want Labor defeated. The dispute does not require a fabricated confrontation or an anonymous source being presented as a named official speaking for the whole party.
For readers, several things remain uncertain. The full electoral effect of Advance's media buy is unknown. The available reporting does not establish how many voters have seen a particular ad, whether exposure has changed their preferences or whether the Coalition would have performed better without it. Future polling or election results may add evidence, but neither can be assumed now. The only sound present-tense claim is that concerns have been raised and the organisation disputes them.
The central question also extends beyond this campaign. As more political groups adopt synthetic imagery, the public must assess not only the message but the medium delivering it. Clear attribution, honest labelling and the distinction between an illustrative attack and a purported real event will become more important. Political disagreement is legitimate; presenting inventions as documentary evidence would be a different matter.
Advance has set out to make opposition to Labor visible in targeted Victorian seats. Some Coalition figures believe its chosen tactics may damage the very result they want, while the group maintains its strategy makes electoral sense. The election will test those rival judgments. Until then, the story is a verified argument about tactics and political advertising, not proof that the campaign has already changed the balance of power.
