1. Extra Categories
  2. Editor's Pick
August 29, 2018updated 04 Jan 2019 4:37pm

Artificial intelligence and the future of politics

By Luke Christou

Trust in the British political system has been falling for decades.

It hit a new low in the United Kingdom in 2017, according to Edelman’s annual trust and credibility survey.  Many feel that life isn’t as fair as it used to be, societal values have been lost, and politicians are at risk of losing control. Just 29% of respondents had trust in the institutions that shape our society.

Following a bitterly fought Brexit campaign, those results are hardly surprising. Depending on which side you listened to in the lead up to the referendum, Brexit would either cause or solve just about any problem that society could face.

The mixing together of fact and fiction has continued to plague Brexit negotiations ahead of the March 2019 deadline, as leave and remain became soft and hard.

As is often the case in politics, point-scoring between parties has been put ahead of the best interests of the people.

However, Michihito Matsuda, a candidate in Tama City, Tokyo’s mayoral elections earlier this year, believes that he has an answer to the problem: artificial intelligence.

As part of his election pledge, Matsuda – who was working alongside Softbank vice president Tetsuzo Matsumoto and former Google vice president Norio Murakami – proposed that decision-making and policy changes should be decided by an artificially intelligent machine, which would use the datasets that the district had at its disposal to make informed decisions that were best for the people of Tama City.

Tokyo wasn’t ready for an AI leader. Matsuda placed third in the election, receiving 4,000 votes. However, with research accelerating far faster than experts thought it would, could AI be the politician of the future?

BrexAIt: Is artificial intelligence the answer?

Brexit has come too soon for artificial intelligence to play a key part in its eventual outcome.

According to 2016 United States presidential candidate Zoltan Istvan, a leading figure in the transhumanist movement, AI is currently only capable of completing basic tasks. However, both funding and capabilities are ever-increasing, which will become increasingly more useful to humankind with each passing year.

“It’s unlikely that AI will provide solutions to Brexit over the next 24 months. If negotiations to Brexit were to take place in 7-10 years, AI might have a significant role,” Istvan told Verdict.

However, AI could still have a role to play in a post-Brexit Britain.

According to accounting firm KPMG, AI could help to circumvent a lack of skilled workers caused by losing access to the EU labour market by relying on robots to perform low-skilled jobs. Likewise, automation would also bring down costs and negate losses caused by raised import and custom fees.

AI in its current form is already capable of taking on basic jobs. In the UK, robots have already taken on 1% of jobs in the manufacturing industry. That climbs to 3% in Germany and 5% in South Korea.

Istvan told Verdict:

“Currently, the not very powerful AI we have today is only useful for basic tasks. But every two years or so, AI is doubling its capacity, and that kind of exponential growth will lead to a machine intelligence that is very useful to humans,”

The roles that AI can take on in the political world are limited. However, Istvan believes that, should AI continue to progress at the current rate, it would fare far better than our current politicians, should Britain decide to reverse its decision in the future.

“If Brexit fails, or the people vote to change their mind in a decade’s time, it might be AI that brings the UK back to the EU.”

Technical difficulties

While AI has massive potential, it still has a long way to go before it is ready to perform complex, difficult tasks. The self-driving car industry is evidence of that. The death of a pedestrian caused by the technology earlier this year showed that AI can often struggle to predict and react to human behaviour.

The use of AI in politics comes with its own set of problems – notably, the ability to act in a fair way, free from biases caused by its programmer or historic data.

Of course, for AI to be used successfully in politics, it would rely on people showing trust in the technology behind it.

That’s a big ask, given the recent misuses of technology in politics.

It was an algorithm that flooded our Facebook feeds with fake news. Bots, autonomous accounts programmed to spread biased opinions far and wide, that give the illusion of widespread support or interest, helped to spread it.

The information that we put into social networks has also used to profile us, pick out those susceptible to influence, and target them ahead of key political votes.

It is widely believed that AI-powered technology was used to manipulate voters in the 2016 United States presidential election, which may or may not have put Donald Trump in the White House. This same method may have also been used ahead of the Brexit referendum and the 2017 United Kingdom general election.

The Cambridge Analytica scandal put that knowledge in the public eye. As a result, most people are now aware of how technology, such as big data and machine learning, can be used to trick and manipulate the public.

“Algorithms are neutral, they’re fair, they’re not human so they’re not biased or prejudice, and they’re the best thing ever for objectivity… In your dreams,” Noel Sharkey, a professor of artificial intelligence and robots, said at FutureFest earlier this year.

Businesses are already using algorithms to make important decisions. Some 33% of employees use AI in the hiring process, while police forces are using AI to determine whether offenders are at risk of reoffending following their release.

The HART system, developed by the Durham Constabulary and a team of computer scientists, uses 34 data points to assign offenders with a risk level.

Reviews found that the predictions made by the machine differed greatly from those made by humans. This, you would think, is due to the removal of biases in decision-making.

However, it isn’t uncommon for human bias to find its way into algorithms and technology.

As well as age, gender and criminal history, the algorithm also used an offender’s postcode to determine their risk level. The system was more likely to give those from poorer areas a higher rating.

A similar system used in the United States would frequently issue black offenders a higher rating. The COMPAS system was twice as likely to incorrectly assign black offenders as high risk, and twice as likely to incorrectly assign white offenders as low risk.

Relying on human input and historic data to form conclusions, AI in politics would undoubtedly face similar issues of bias in its current state, whether due to the ideologies of its creator, or the data that it relies on.

However, leading AI tech companies, such as IBM, are working hard to overcome this problem. Funnily enough, they’re using AI to do so.

According to IBM Watson Chief Architect Ruchir Puri:

“We [IBM] have several ongoing projects to address dataset bias in facial analysis – including not only gender and skin type, but also bias related to age groups, ethnicities, and factors such as pose, illumination, resolution, expression, and decoration.

“More broadly, we are also developing algorithms for detecting, rating, and correcting bias and discrimination,”

Rapid advancement

While some worry that AI could eventually become a force of evil, transhumanists are convinced that advancing technology can only help to improve the human existence.

“In the next 10 to 20 years, we will have robots with AI that are literally as smart as humans,

“The only difference is they won’t be able to lie if they’re programmed against it. Therefore, if they were to lead our governments, we would actually be able to trust them to always do the right thing,”

Perhaps some of the ideas that Istvan and the Transhumanist Party champion, the eventual ability to overcome ageing for example, seem outlandish and unachievable. However, it’s difficult to fault his view on AI.

According to industry experts, AI will outperform humans at writing essays by 2024 and write a bestselling book by 2049. By 2053, it will be a competent surgeon. By 2062, machines will be better than humans at more or less everything.

With AI technology developing far faster than experts thought it would, it seems like only so long before a robot is running the White House.

Read more: No killer robot ban a “danger to humanity” warns Noel Sharkey