We Have Arrived at the Intersection of Stupid and Smart
Politics are getting dumber. AI is getting smarter. What happens when they meet?
Today’s Multimodal Large Language Models are incredibly smart, and are available for use by billions of people worldwide creating the potential to make the whole world a smarter place. On the other hand, politics in the United States seems to get stupider and stupider every day. What’s going to happen when we get to the intersection of Smart and Stupid? Can US voters turn off Stupid Street and turn on Smart?
Stupid Street
The descent of US politics into greater and greater stupidity seems precipitous. The lefties have adopted extreme “wokeness,” spending their time on political and gender issues that fail to resonate with mainstream America. Radical lefties on college campuses restrict free speech by preventing those who disagree with them from speaking their minds — now manifesting itself in the form of campus disruptions disguised as protests.
Much worse is the descent of MAGA Republicans down the path of anti-intellectualism, disrespect for experts, book burning, theocracy, and dictatorship. Right-wing conspiracy theorists embrace anecdotes rather than the scientific method, and worse yet, don’t understand the difference. Their flag-bearer, Donald Trump, doesn't read books and is uninterested in geography, world history or the nuances of military strategy. As George Santayana said, “Those who cannot remember the past are condemned to repeat it.”
In US society, misinformation is rampant in the domains of public health, climate change, economics, election integrity, immigration, and healthcare. In each of these areas, taking an informed position on policy requires knowledge of the facts and the ability to reason using those facts.
Determining the facts on these issues requires study of peer-reviewed articles, reading of reports from reputable organizations, watching unbiased news sources, and objective personal observation. These things are difficult for the average voter, partially because of time and interest, but also because of never having learned the necessary research techniques. What is a peer-reviewed publication? Which are reputable organizations and news sources? What is an objective personal observation? Only 35% of Americans have completed a Bachelor's Degree which, in the United States, is the educational level where these research techniques are taught.
Smart Street
Adoption of the Internet and use of web browsers and Google Search made the breadth of human knowledge available to the world’s connected population.
In 2021 Google DeepMind’s AlphaFold 2 cracked a protein folding problem that had stumped scientists for decades. By predicting the 3D structures of proteins, AlphaFold 2 opened new doors for understanding diseases, designing drugs, and unlocking the mysteries of life itself.
On November 30, 2022, OpenAI released ChatGPT, a large language model (LLM) that enables users to converse with an AI Assistant in most human languages. Monday May 13 of this year, OpenAI announced their new Multimodal LLM GPT-4o and demonstrated its native voice and vision capabilities.
This model joins Claude, Gemini, Llama, and Grok giving US voters access to a variety of free or minimally expensive AI Assistants with more and more becoming available every day. While these models are fine tuned and otherwise restricted from giving direct election advice, they are still valuable tools for candidate and issues-related research and dialog.
Voters can use Internet-connected AI Assistants to access peer-reviewed journals, government and educational institution websites, and news organizations that adhere to strict journalistic standards. Voters who want to talk through their reasoning on various policy issues can use one or multiple AI Assistants as a smart, willing companions for discussions about election-related issues 24/7.
Arriving at the Intersection
The voters currently exhibiting stupid behavior will increasingly be exposed to smart technology. Having access to the Internet via today’s Google or Bing gives voters access to an overwhelming amount of information. But before using it to make voting decisions, voters should validate its accuracy and identify any bias in reporting.
Since most voters don’t have these validation skills, or the time to use them, tomorrow’s AI Assistants will do it for them. In the future, your AI Assistant will explain how it makes sure what it is telling you is correct, and it will encourage you to validate information you gather from other sources.
Many US voters either lack reasoning skills, or the discipline to use them, or both. Inductive reasoning starts with specific observations and forms general theories. Deductive reasoning starts with general premises and forms specific conclusions. Our future AI Assistants will help us apply these reasoning processes, allowing voters to use these techniques when evaluating candidates and issues.
Busy, distracted, and stupid voters tend to accept information and propositions that confirm their biases. More thoughtful, focused, and smarter voters learn to question their own assumptions and explore alternative explanations and options. AI Assistants will help voters learn to question purported facts and will help them seek objective truths.
AI Assistants that are smart, uncensored, and unbiased, have the potential to help voters improve the quality of political argument through discussion and reasoning. AI Assistants will encourage learning by helping voters identify relevant facts, evaluate those facts, and then reason with those facts to reach conclusions about how they should vote given their own personal values and objectives.
Turn on to Smart Street
Turning off Stupid Street will require both courage and technology. We will need courage to accept change, courage to recognize our current lack of knowledge. Then it will require courage to use AI Assistants to seek a more fact-based and analytic approach to our voting decisions.
Which candidates should we support? How should we vote? To help us with these important decisions, our technology will behave like a teacher by helping us curate our information stream, and apply reasoning skills to our decisions.
The lessons for AI developers and entrepreneurs to learn are:
- Providing voters with accurate information is not good enough.
- Using reasoning and logic while providing information to voters is not good enough.
- AI Assistants need to teach voters how to check facts against validated sources and then help voters apply proven reasoning techniques to arrive at their own smart voting decisions.
An ideal AI Assistant will help voters absorb these lessons and use them in both their connected and in their unconnected lives.
Postscript: Since writing this article I learned more about the Kahn Academy and how its Founder and CEO is using AI to educate children. His educational AI bots are designed to not just teach facts, but also to teach how to evaluate facts, how to learn, and how to reason. See his book Brave New Words: How AI Will Revolutionize Education (and Why That’s a Good Thing).
Question for the Reader: What are the ethics of using AI to make us Better Voters?
Even I have some questions about the recommendations I made above. I would like to hear your thoughts.
- Can we trust AI to teach voters how to discern facts from misinformation?
- Can we trust AI to help voters use basic reasoning and logic?
- Is it right to use AI to modify the behavior of millions of voters? Is it right if it makes our voting decisions better reflect our own personal principles and values?
What do you think? Please leave a comment.
