Over the years there have been many risks associated with artificial intelligence, some of them more legitimate than others, but several focus on the processing capability of artificial intelligence technologies and how they significantly outmatch anything we humans could ever achieve biologically. One of the ways in which this could become a significant problem is with the AI to meddle in a country’s state or national elections.
With the threat of election meddling now looming over nearly every country in the world thanks in part to our ever-increasing connectivity, naturally questions will be raised about the potential role artificial intelligence technologies could play in such events.
Unfortunately for us, AI could potentially throw challenges that make things such as being able to tell the difference between a real video and one that’s been manipulated to show or say something completely different, incredibly difficult.
However, all is not lost. AI technologies could also be used to help make meddling in another sovereign nation’s election process extremely difficult, too. AI could be used to secure and protect voting machines, detect fake news and content, and understand the behaviors of hackers who could look to subvert democracy in this way.
Understanding the Threats
When it comes to behavioral prediction, AI will be capable of collecting, processing, and interpreting vastly more data than human analysts and will be capable of learning to differentiate between different types of traffic in an effort to distinguish regular users from malicious hackers.
This could be a particularly useful capability in the cyber security systems used to protect voting machines. Understanding when a breach of security had occurred is one thing but having a security system that is aware of the type of traffic being received and being able to gauge the risk it poses will significantly enhance security.
There are also the challenges posed by new AI technologies in areas such as personal assistants. Siri and Cortana, for example, have fairly generic responses to questions regarding user politics, but newer systems will be able to learn a person’s political beliefs and offer advice that caters to their positions on various different issues.
There is understandably concern that these systems will be vulnerable to attacks from rogue hackers or well-resourced nation states and the same principles of cyber security will apply to these technologies. Artificial intelligence could also be used to ensure that this data stayed protected.
Decoding Fake News
When it comes to meddling in elections, it doesn’t just down to manipulating voting machines. Flooding the web with inaccurate and politically bias information and ‘news’ is one way in which those that may wish to influence an election could attempt to do so. There has been much attention recently on the issue of fake news and various methods of dealing with it have been proposed by many different groups working on solutions.
Machine learning is a key branch of AI in this area as these systems would look to understand the slight nuances that often separate fake news from real, legitimate reporting and could also be built into apps that help users identify fake news stories or the sites known for publishing them.
By focusing on aspects such as the wording of an article, use of blue language, interpretation, publisher credibility and so on, machine learning algorithms could alert readers as to how likely it is that the article they are reading is trustworthy or potentially fake news.
Understanding linguistics has become a huge part of being able to detect patterns in fake news stories and by focusing in on tiny details such as those mentioned above, coupled with other basic checks the user could even do themselves such as whether it is traceable via an author can help to significantly reduce the impact these stories have on official elections and the public themselves.
Into the Future
With artificial intelligence technologies becoming increasingly more widespread as well as steadily becoming more powerful and capable, steps will certainly need to be taken to ensure that they do not then readily enable even more advanced kinds of attacks on democracy such as the meddling in an election process.
Malicious hackers, foreign saboteurs, and those that wish to undermine the will of the people will all try to utilize artificial intelligence to their own ends, and what’s unfortunately highly likely is that this will mean using AI to attempt to manipulate election results. Developers of artificial intelligence technologies will need to pay particular attention to this issue and try to work out a way in which these technologies are more focused towards preventing such activities than partaking in them.
Artificial intelligence, and branches such as machine learning and deep learning will all have a part to play in the fight against election meddling. How well they are utilized in protecting against all out, AI-powered cyber warfare that aims to dethrone democracy will depend on how seriously this threat is taken. Fortunately, it appears as if we are beginning to realize the urgency of this threat and starting to take steps towards tackling it.