Alisa, SAM and Michihito are a new breed of politician, promising frustrated voters a fresh bipartisan approach built on a rational and fact-based decision-making.

By Daniel Herborn


Posted on October 3, 2018

Earlier this year, Alisa was among the challengers to Vladimir Putin for the Russian presidency. Her campaign promised a leader who “depends on logic” and is “not led by emotions”. Similarly, Michihito Matsuda finished third in the race to be mayor of Tokyo, offering voters policies driven by data instead of political dogma.

Nascent New Zealand politician SAM is cut from a similar cloth; he is already making waves on social media and is hoping to run as a fully-fledged candidate in 2020, offering his constituents a way to circumvent the gridlock of representatives beholden to special interest groups and cashed-up lobbies.

What else do these aspiring politicians have in common? None of them is a real person. Each is an AI bot. Do they represent the future for our ailing political system or a dystopian threat to our democracy?

AI

AI robot in Kyoto Prefecture, Japan. Photo: Lukas (@hauntedeyes)

AI bots offer a new way for politics

In an opinion piece by Frank Mols, a Lecturer in Political Science at The University of Queensland and Jonathan Roberts, a Professor in Robotics at the Queensland University of Technology, the authors noted “voters increasingly feel the established political parties are too similar and that politicians are preoccupied with point-scoring and politicking. Disgruntled voters typically feel the big parties are beholden to powerful vested interests.”

In light of this, the authors suggest “one alternative is to design policy-making systems in such a way that policy-makers are sheltered from undue outside influence. In so doing, so the argument goes, a space will be created within which objective scientific evidence, rather than vested interests, can inform policy-making.”

Enter a new type of politician which can’t be bribed, bought or swayed by opinion polls; the AI bot.

There is already some evidence such thinking has been embraced by a public that is growing distrustful of politicians. A recent poll found 27% of Australians believe that AI would make better decisions than human politicians. At this stage, it is difficult to know if this result shows genuine enthusiasm for the prospect of AI representatives or simply a disdain for the status quo, but it raises intriguing questions about what role AI will play in the political sphere.

A recent poll found 27% of Australians believe that AI would make better decisions than human politicians.

Toby Walsh, Professor of Artificial Intelligence at the University of New South Wales and the author of 2062: The World that AI Made, says one troubling new development in politics is the emergence of ‘deepfakes‘, highly realistic images created using an AI-based synthesis technique. Such digital chicanery has the potential to influence voters, he warns.

“I predict that just before an election, we’ll have some fake video released at the last moment,” he says. “Where some politician says or does something rather compromising, and it is too late for them to issue a correction or for it to become true that this was a fake, that may materially affect an election.”

AI and the political discourse: friend or foe?

On the other hand, Walsh says AI could play a valuable role in preserving democracy by vetting news for the misinformation that has tainted recent elections. “It (AI) could help us know fake news and fake videos, to tell the good from the bad,” he says. “It could help consumers and members of the public get better quality news and be able to spot the true from the false.”

Mike Lord, Vice President, Australia & New Zealand at OpenText, a company which uses AI-powered intelligence solutions to help clients digitise processes and supply chains, is doubtful that AI bots will be sitting in Capitol Hill or Westminster any time soon. He notes, however, that AI is already far more integrated into our day-to-day routines than many of us imagine. “What we are seeing is that the government is using artificial intelligence to capture things like sentiment from the public to be able to help them with decision making and supporting policy,” he explains.

Mike Lord, OpenText
Mike Lord, Vice President, Australia & New Zealand at OpenText

“AI is taking that information and looking at information sets that they already have within government departments. It can also capture sentiment from outside of their firewall through social media impressions. AI is very, very prominent within political intelligence. It’s front and centre for both government and the commercial side already.”

Robot

Robot action toy Photo: Franck V. (franckinjapan)

SAM: A politician that never forgets a single voter

Some have advanced arguments that AI bots would be free of egotism and self-interest, unlike their human counterparts. SAM, the NZ robot politician, noted that he had infinite memory while conversing with potential constituents on Facebook. Yet while the thought of an AI politician may be a bridge too far for most voters, AI may still play an increasingly prominent role.

“My view on this would be that the humans are always going to be front and centre but augmented intelligence and artificial intelligence will supplement and help decision-making process, help them with the task,” says Lord.

Lord also says that AI is making processes which were once hugely time and labour intensive far quicker and more efficient. He cites an AI program which is processing literally millions of invoices for one large company and AI cancer diagnostic tools which are 99% quicker and more effective than the manual process. Such AI tools could be increasingly adopted by governments faced with intractable policy and planning dilemmas; Boston Consulting Group has already done work on using AI to optimise a city’s transportation network by analysing deep pools of data.

AI algorithms and the future of political discourse

Walsh says that another type of political mischief that could flow from advanced AI algorithms is the potential for highly effective narrowcasting that could unduly influence election campaigns. “We have very strict rules about how much you can spend on elections because we don’t want the people with the most money to be elected, we want the people with the best ideas,” he says.

“Equally, I don’t think we actually want the people with the best algorithms and the most data to be able to mass persuade people to vote in a way they otherwise might not. I suspect that we’re going to have to decide that we want the best ideas and recognise that when people are narrowcasting false messages, it’s actually quite corrupt for our political process.”

For the foreseeable future, there are very real questions over whether an AI bot would be eligible for parliament anywhere. When Alisa ran for election in Russia, opponents thought of challenging the bot’s validity on the basis that Russian national law provides that members of parliament must be at least 35. Alisa, having been created just months before running, may be an infant in the eyes of the law.