Generally Productive Talks #8 - Ai and Politics

Introduction

Previous blog posts have been easily linked to Design Development, today's is not on the surface, however, given its current and imminent impact I felt it the right topic to add

Ai has and will change so many areas that influence our lives, it will do the same with Politics, we should now begin the work to plan for its effects and mitigate its dangers

The Past

The first use of anything related to Ai in Politics can be traced back to the 1960s when computer modeling was first used to predict the US presidential election, where the University of Michigan's UNIVAC I correctly predicted JFK defeating Nixon

Its use since then remained similar until the 2008 US election when the Obama Campaign used data analytics and machine learning to identify voter groups to target, this developed further in the 2016 US election where the Cambridge Analytica scandal and its impacts brought the practice to public attention. The result of this was a wake-up call to democracy's around the world and one by one they began putting regulations into place to add transparency and ethical practices

Ai continues to be used in elections with algorithms analyzing voter habits, and targeting marketing on specific groups, along with the adoption of 'social bots' used on social media to perform certain actions like liking or sharing posts, following users, and engaging in conversations on specific topics, this can be particularly dangerous as they can spread disinformation or amplify specific messaging to drive public opinion through the shear volume of information, this has already been seen in US and Brazilian elections but is clear to have been used more widely and by domestic and international interests

Today

Ai's use in Politics is expanding rapidly and its uses are increasing:

• Campaign Management - optimizing message avenues for revenue spend

• Predictive Analysis - helping determine how to target swing voters

• Speeches - analyze and tailor wording in speeches for specific audiences

• Social Media Monitoring

• Policy Analysis

All of the above can help better run campaigns and speak to people in more relatable ways, however, it can also be used to trick and convince people of things that may or may not be beneficial or even true, it all depends on the motives of those using the technology

One large danger in how Ai is affecting politics today is the algorithmic dominance in what people consume, apps and websites designed to encourage constant use through matching desired content to a user, the danger of this is that the user never becomes challenged in their views regardless of what they may be and is only reinforced in whatever they prefer as being the right things to like, follow and consume more of. This is resulting in more polarisation of beliefs, less engagement, and discussion, and raises the chance of conflict as early-stage discussion never occurs between different sides and when two sides do meet, both are so entrenched in their own mindset that dialogue is difficult and conflict becomes easy, a truth that is becoming more prevalent each day

The Future

Ai and Politics is an active subject today, but in the future, it will only become a more active partnership, below are possible things to expect in the coming years:

5 Years

• Voter Analytics - more targeting by increased amounts of information mining

• Misinformation - will both increase in the amount generated but also in the amount detected and removed, it will be a constant fight between generators and providers

• Administration - Ai will take over many human-operated activities in registration, counting, and result reporting but must have safeguards to isolate from outside forces

10 Years

• Ai Assistants - real-time analysis through assistant bots both for politicians and reporters making all things more streamlined and regimented/controlled

• Ai-Generated Content - most content will be produced via Ai with articles, ads, social media posts, and videos created for targeting specific voters

• Ai Assisted Decision Making - decisions will be based largely on the findings from algorithms and Ai programs

20 Years

• Direct Democracy - with Ai powering many actions currently taken by humans (including many jobs) people will have more time and democratic politics could revert back to its roots with Athenian-style direct democracy with each person able to have direct power

• Ai Powered Regulation and Monitoring - policing disinformation or hate speech and ensuring rhetoric doesn't become too heated, it may be possible to have a more civil discourse than currently possible online, but could also go to a degree where differing opinion is impossible to state which raises its own concerns

• Ai Generated Candidates - with enough time and data it could get to a stage where Ai can handle the entire process and produce candidates that would be more appealing than flesh and blood alternatives, at which point, lines become blurred

There are many more possibilities, benefits, and issues to raise, each political system will be affected differently, but 1000 words are not enough words to explore everything...

My thoughts:

You cannot uninvent something, once it is done, it is available and will be used. Also, you cannot blame a tool for the damage it does, rather it is humanities shortcomings that make the technology problematic

Ai will be used in politics, it will be used in both good and bad ways with the bad ways becoming increasingly problematic and difficult to manage unless drastic actions are taken early to first dissuade people from using it in any way that can cause trouble (regardless of their political leanings) and steps are taken in financing, live traceability, and content that will prevent outside forces being able to promote their own interests

To do the above though will mean more laws, fewer freedoms, more red tape, and restricting actions that work for political parties that want to win, as such I do not believe politicians will do what needs to be done and things will inevitably get messier, what happens after that?

Outlook is not good...