ChatGPT and automated propaganda
Now that language A.I. is able to join online political debates and thereby influence the actions of governments, limiting government powers has never been more important.
The chatbot ChatGPT, launched by OpenAI in November, showcases that A.I. is now on par with activists when it comes to flooding the media with biased, nonfactual articles and posts that make the government waste money and violate our liberties. This has led to calls to ban or control the A.I., but would that really solve anything? Instead of banning A.I., we should be reforming the political system to stop "collective" demands, whether they are worded by human or electronic drones, from interfering with businesses or individual lives. In fact, A.I. could help us do this if we employed it as a tool for liberty.
ChatGPT heralds a new era. It impressed people around the world with its ability to write detailed, human-like texts, from serious college-level essays to jokes like this biblical manual on removing a peanut butter sandwich from a VCR.
Nonetheless, critical voices highlighted its lack of factual accuracy. The programming platform Stack Overflow even banned the software on the grounds that it produced answers to programming questions that sounded convincing but turned out to be false.
New-generation language A.I.s are a powerful political tool. A simple way to use them would be to support the writing of articles that you then polish and publish in your own name. But this is child's play compared to its true potential, which lies in automation. Since the instructions for chatbots are just text data that themselves can be automated, the A.I. can be set up to "contribute" biased posts or articles to online discussions. And it can do it so much faster than people. In minutes, it can write two separate posts posing as different people that arrive at the same political conclusions by different routes. It could create debates on any chosen subject on demand.
This of course has resulted in fears regarding political misuse. But most of these fears come from people who dominate online debate right now and see their position threatened.
A chatbot's connection to reality will always be indirect at best. Ask it a programming question like the guys at Stack Overflow, and it becomes obvious that the A.I. doesn't program itself; it just mimics reasonable speech. Is political dialogue at all different nowadays? Look at what current political debate draws on for the most part: previous political debate. Online political discourse is already in a feedback loop — one that takes little to no input from the everyday lives of real people and the problems they face. Apocalyptic climate change alarmism ruminates ad infinitum while families unable to pay gas for their car because of emission targets and energy taxes remain unheard — or get reflexively and repeatedly told to use public transport even though they live in the middle of nowhere. Engaging in political discussions online is already a lot like talking to a very bad chatbot. Throwing ChatGPT into the mix is really just the next logical step.
So we can be thankful to chatbots for making the problem abundantly clear: politics today is driven by mindless online mobs.
As bots gear up to join those mobs, we should be looking for ways to bolster safeguards around fundamental rights — like the rights to private property and free speech. Furthermore, we should be looking for ways to limit the government's ever-ballooning budget.
Regardless of whether the mob is led by human activists or artificial intelligence, the real danger is that the Constitution is vulnerable to the whims of a majority.
Ironically, the chatbots are our best shot at achieving this. Instead of trying to ban the A.I., conservatives should prepare to fight fire with fire, using chatbots to mass-produce articles and posts that emphasize individuality and entrepreneurship, demand constitutional reform, and push back on every collectivist and technocratic post in the media.
A.I. is coming to political discourse whether we like it or not. Since enemies of freedom will use it anyway, why shouldn't we?
Torben Halbe is a contributor for Young Voices, non-fiction author, and liberty activist at the EGO Institute based in Berlin, Germany. He holds an M.Sc. in biology from ETH Zurich. Follow him on Twitter: https://twitter.com/TorbenHalbe.