New Zealand needs to wake up to the threat of online voter manipulation in this year's election, experts say.
Some think the country should consider following Europe's move to ban microtargeting - which exploits your social media profile.
"It is, I must stress, going to be a campaign and election that is historically unprecedented," Disinformation Project researcher Dr Sanjana Hattotuwa said.
Twitter, Google, Facebook and Apple have each introduced tighter policies on campaign advertising since the Cambridge Analytica scandal in 2016.
But these policies do not apply to the data brokers that the tech giants still sell billions of bits of highly personal data to; and the brokers were still used by political consultants to create vivid portraits of what pushes the buttons of an individual or group of voters.
Joshua Ferrer, who wrote a 2020 report into online campaigning in New Zealand for Transparency International, was more than wary.
"We haven't had a Cambridge Analytica moment, we haven't had a Russia interference moment," he said.
"But that doesn't mean that the threat isn't real.
"The technology is there for one of those moments to happen."
His 2020 report found the guardrails around political online campaigning were weak or missing here.
This at a time when the threat was becoming more "sophisticated", and as tactics honed overseas were imported, according to Internet New Zealand.
InternetNZ has sounded warnings about online interference following the last two elections, and in 2020 said: "Over the past two years, our concerns have become more pressing, but as far as we know remain unaddressed."
At the same time, the price of interfering was only going down.
A thousand dollars could buy a lot of very aggressive, one-eyed messaging pumped into social media echo chambers, the experts say - whether you were a political party, a party's proxy, an anti-vaxxer, or Russia.
It was "possible to test thousands of message variations, assessing how each individual responds to them, and changing the content in real time and across media in order to target and retarget specific voters", a study said.
It was debated globally just how effective this was and been suggested pscyhological microtargeting could actually help politics.
A 2022 US MIT study said its results "suggest that political microtargeting can confer a sizeable persuasive advantage over more traditional messaging strategies".
And where campaigners of whatever stripe were looking for an edge, and do not have to reveal just what they were doing - as was largely the case in New Zealand - well, why wouldn't you?
'Profoundly disturbing'
Hattotuwa has deep fears, spawned from years looking at the growing use of disruptive campaign data since 2010 in Sri Lanka, since 2020 in New Zealand most recently around Covid and since early 2022 to do with Russia's misinformation campaigns locally centred on its war against Ukraine.
"What we have been studying about what's been going on and what went wrong ... is going to impact the democratic fabric and electoral integrity for all elections henceforth," he said.
"It is profoundly disturbing.
"It is extremely dangerous. It is highly divisive, and it is intended to go to the heart of electoral integrity. That is not a future projection, a hypothetical future scenario, that is what is present, pulsating, expanding and entrenched."
Victoria University professor Jack Vowles helped with research that found some online half-truths in 2020's election, but little fake news.
Microtargeting was a valid tool that had been used for years, such as using the electoral roll to target teachers with education policy brochures in the mail, so banning it would be "draconian", Vowles said.
It was only manipulation that was the danger - though whether it even worked was another question.
"There is clearly evidence that people try to do these things.
"But overall, looking at the results of an election in aggregate, it's not clear that this kind of thing has a great deal of effect", though determining the effects was difficult, Vowles said.
InternetNZ's senior policy advisor Michael Daubs said it was speculative to talk of impacts on the election at this distance.
But like Hattotuwa and Ferrer, he was certain New Zealand was behind the eight-ball.
"We're not necessarily ready to deal with an onslaught of misinformation in relation to the election," Daubs said.
"What we're really looking for is that, when an ad is delivered, because these ads can be microtargeted, that individual person at that moment needs to understand exactly where that message is coming from."
Reports by the Electoral Commission after the vote were no help, he said.
"This is about making decisions when someone goes to vote."
Follow Europe?
Ferrer asked if it was good enough for Europe to act now, to curtail microtargeting and resist interference in the 2024 EU parliament elections, what about New Zealand in 2023?
"We know the parties are spending lots of money using them [voter targeting tools].
"Now is the time before a huge crisis of confidence occurs to do something about it. I think the European Union has taken a tremendous step forward.
"And I think it's time for us to, as well," Ferrer said from California, where he lives.
The Electoral Commission said prohibiting microtargeting would require legislative change.
All election advertisements, in all mediums, had to include a promoter statement that tells people who's behind the ad, it said.
There were also spending limits for parties, candidates and third parties in the run-up to an election, and rules around when election advertisements can be broadcast on television and radio.
Researcher dismisses fears
A researcher into social media political messaging in New Zealand says there is no reason to worry about voters being manipulated online.
Dr Mona Krewel of University of Victoria runs a project measuring fake news, conspiracy theories and half truths in political communications of parties in social media.
This was not a huge issue at the 2020 election, Krewel said today.
"At least with regards to the established mainstream parties, fake news, conspiracy theories and half-truths are still not a huge issue ... and there is no reason to worry, as my current data shows.
"Mis- and disinformation in New Zealand are almost exclusively used by certain fringe parties ... and interest groups ... or anti-vaxxer influencers ... and not much beyond their circles."
Krewel said while their study did not look at microtargeting in depth, any targeting that did show up during a month of monitoring in the run-up to the 2020 election did not seem very elaborate and was "fairly standard practice".
The degree of microtargeting was the key, she said.
"It still seems rather unproblematic to target people on broad group characteristics such as regions in which they live or age-groups they belong to.
"It ... gets more problematic once you allow real microtargeting based on more sensitive information such as one's sexual orientation or health status or psychological traits, and it is an ethical question whether political ads should be allowed to profile people based on having revealed this information about themselves for completely different purposes."
This latter use of a person's online profile, gathered by harvesting what they did and said online, analysing and processing it, has been widespread in the US, according to many media reports.
Krewel, though, had doubts how effective this was.
"Methodologically, it is not the magic trick that allows you to manipulate people and win elections," she said.
"Voters are not puppets just because you have some information about them."
The EU's move to curtail political microtargeting might promote transparency, limit misuse of sensitive data and help prevent foreign interference, but people who wanted to use a certain social media service would still sign up to terms and conditions that allowed their data to be used for political advertising, regardless, Krewel said.