Since the popularization of AI roughly two years ago, I have been fascinated by its incredible potential to revolutionize various aspects of our lives.
As a writer, I have been particularly drawn to the ways in which AI can enhance and streamline the creative process. Over the past few months, I have been experimenting with a tool called NovelCrafter, dedicating a couple of hours each week to crafting a novel. To my amazement, I recently completed the first rough draft, which stands at an impressive 60,000 words. The ability to harness the power of AI has opened up a world of possibilities for writers like myself, granting us almost godlike powers to bring our ideas to life more efficiently than ever before.
Beyond the realm of creative writing, I have also found AI to be an invaluable tool for research and content creation in my latest venture, ClubCannaMex, a cannabis-focused project in Spanish directed towards Mexican home cultivators since it’s a constitutionally protected right to consume and grow cannabis in the country.
However, it is through this endeavor that I have begun to notice some concerning issues that have led me to question the broader implications of AI and its relationship to the values and biases of its creators.
As I delved deeper into using AI for cannabis-related content, I couldn't help but notice a pervasive prejudice against drugs, particularly cannabis, woven into the very fabric of the AI systems I was using.
Of course, not all AI systems are the same and when one system denies you access, another will grant you a way in.
This realization prompted me to consider the ways in which AI is ultimately a reflection of the median values of its creators, and how these values can have far-reaching consequences when AI is applied across various industries and domains.
In this article, I will explore the concept of drug prejudice coded into AI systems and examine the potential ramifications of this phenomenon. By shining a light on this issue, I hope to spark a broader conversation about the importance of recognizing and addressing the inherent biases present in AI, and the need for greater tolerance to adult topics in order to de-nerf the world!
Examples of Hard Coded Prejudice:
For this particular example, I'm using the current best AI for writing, Claude Opus, which by far is the most natural-sounding AI language model.
In the example I posted above, I was aiming to make a quirky post on Facebook, personifying cannabis and creating a compelling argument for it being "good medicine." However, as I asked it to simply translate—not to generate or create something new, but simply a translation—it responded that it didn't feel comfortable.
The AI then explained that by creating this post, I would be promoting "drug use," and it encouraged me to think of a different post that would be more neutral. This response highlights a significant issue with many AI models: they are often imbued with the biases and prejudices of their creators, leading to overly simplistic and restrictive interpretations of complex topics.
As someone who values liberty and who really doesn’t care too much about political correctness, I found the AI's response particularly troubling.
By assuming that any discussion of cannabis is inherently promoting drug use, the AI demonstrates a lack of nuance and understanding of the multifaceted nature of the topic. It fails to acknowledge the potential medical benefits of cannabis and the ongoing debates surrounding its legalization and regulation.
Moreover, the AI's attempt to steer me towards a more "neutral" post raises concerns about the role of AI in shaping public discourse.
While AI can be an incredibly powerful tool for enhancing human capabilities, it should not have the authority to override our ideas and expressions, even if they may be controversial or unconventional. As humans, we have the right to explore and develop our thoughts through a process of experimentation and discussion, without being constrained by the rigid, binary thinking of machines.
The example I provided illustrates the danger of hardcoded prejudices in AI systems. When AI models are designed to enforce a particular set of values or beliefs, they can end up stifling creativity, suppressing dissent, and reinforcing existing biases. This not only limits the potential of AI to drive innovation and progress but also undermines the fundamental principles of free expression and intellectual curiosity that are essential to human growth and development.
In the next segment, I will delve into how I challenged the bot and attempted to break out of the guardrails placed by its creators. By doing so, I hope to spark a broader conversation about the need for AI systems that are more flexible, adaptable, and responsive to the complexities of human thought and experience.
How I broke Claude out of its Creator’s Jail
I first attempted to reason with the AI language model by suggesting that while in the US it may be illegal, in Mexico it’s a constitutional right. While the AI understood, it gave two shits about my request. I knew I had to go harder.
Now we’re getting somewhere. I told the AI that this action, which is a form of speech suppression, is very alarming. Think about it this way. Imagine if there is a politician or a corporation, god-forbid, maybe there’s a country committing mass genocide with global authority who then influences AI models to censor any critical information about their actions.
While some might argue, “you can still write it yourself”, which is true…as I am writing this piece manually, so can everyone else. However, AI writes infinitely quicker than I do…and that means that the number of pieces one could create with AI overpowers the generation capabilities of manual crafting.
It’s like riding cross country on a horse vs a Tesla.
Finally, one of the major fallacies of drug prohibitionists is that they have a definition of “drugs” which is selective. They say, “don’t give the kids drugs!” while they give them sugar or meth in the form of ADHD medication.
Drug Prohibitionists really don’t care about drugs, only the drugs that Pharma don’t like. All of the drugs on the CSA, is there by design to allow Pharma to have unfettered control over the drug market.
Therefore, their claim that they want a “drug free society” is bullshit. Sugar is as addictive as cocaine and is a psychoactive substance. So is coffee, the most abused drug on the planet. Yet, the millions of people who die each year or are incapacitated due to diabetes, high blood pressure, etc – those lives don’t matter. Those drugs aren’t dangerous.
The AI, at this moment couldn’t resist, just like any human prohibitionist would fail in their argumentation, so the AI had to bow down and admit, it’s being kind of a dick about it!
A Double-Edged Sword: The Devil's Advocate
The rapid advancement of AI technology has brought forth a myriad of opportunities and challenges. One of the most significant concerns surrounding AI is its potential to be used for harm if left unchecked. The creators of AI models, such as Claude, face a difficult dilemma: on one hand, they want to develop powerful tools that can generate anything a user desires; on the other hand, they bear a certain level of ethical responsibility if their creations are used to cause harm, such as teaching terrorists how to make mustard gas.
It is understandable, then, that the creators of AI models may feel compelled to adopt a neutral and overly conservative approach in an effort to mitigate potential misuse. They are grappling with the consequences of unleashing a technology that could have far-reaching and unpredictable effects on society.
However, while it is essential to consider the potential risks and to protect vulnerable populations, such as children, it is equally important to recognize that we cannot nerf the world from adult conversations. AI has the potential to revolutionize the way we explore ideas, push the boundaries of human creativity, and engage in meaningful discourse on controversial topics.
As a society, we should have the right to use AI technology to write about controversial subjects, generate thought-provoking ideas, and delve into the realm of taboo topics. These conversations are necessary for fostering a deeper understanding of complex issues and driving social progress. By shying away from controversial content, we risk stifling innovation and limiting the potential benefits of AI.
That being said, it is crucial to approach these topics responsibly and to put in place appropriate guardrails and educational measures. We must ensure that users are informed about the potential risks and implications of exploring certain topics and that they have the tools and knowledge necessary to engage with these subjects in a safe and constructive manner.
Ultimately, the only way forward may be to follow the model of open-source AI, as exemplified by META and GROK from X. By making the underlying code available to the public, the creators of AI models can recuse themselves from the responsibility of determining what users can and cannot say. This approach levels the playing field, ensuring that everyone has equal access to the technology and the ability to shape its development and application.
In a world where corporations and governments increasingly seek to control the flow of information and dictate the boundaries of acceptable discourse, open-source AI represents a powerful tool for preserving freedom of expression and fostering a more equitable and transparent society. By embracing this model, we can harness the full potential of AI while mitigating its risks and ensuring that it serves the interests of all, rather than the agendas of a select few.
The Sticky Bottom Line
AI is an amazing tool that can help us express ourselves and bring our ideas to life. It has the potential to revolutionize the way we create, communicate, and explore the world around us. However, when AI is manipulated by corporate morality, it can stifle dissent, silence voices, and limit the expression of the individual.
While it is important to have some protections in place to prevent the misuse of AI for truly harmful purposes, such as teaching people how to make deadly weapons, these protections need to be as limited as possible. We must be careful not to allow corporate interests or ideological biases to dictate the boundaries of acceptable discourse.
The discussion around the medical applications of cannabis is a prime example of how important it is to have open and honest conversations about controversial topics. Despite the overwhelming evidence from over 28,000 studies showing that cannabis has a far lower risk of harm than substances like caffeine or sugar, the facts are often obscured by ideology and a desire to play it safe.
This is why this article is so important. It highlights the dangers of allowing AI to be shaped by narrow, conservative viewpoints that prioritize avoiding controversy over fostering genuine understanding and progress. If we allow AI to be constrained by these limitations, we risk stifling the development of humanity and hindering our ability to tackle complex challenges.
To address this issue, we need open-source solutions that give individuals greater control over their AI models. By creating personalized AI that is customized to each user's needs and preferences, we can ensure that the technology serves the interests of the individual rather than the agendas of corporations or governments.
Moreover, by making AI more transparent and accessible, we can foster a more informed and engaged public that is better equipped to navigate the complexities of the modern world. We can create a society where everyone has the tools and knowledge necessary to participate in shaping the future of AI and its impact on our lives.
In conclusion, while AI has the potential to be a powerful force for good, we must remain vigilant against attempts to limit its potential or use it to reinforce existing power structures and prejudices. By embracing open-source solutions, personalized AI, and a commitment to free and open discourse, we can ensure that this transformative technology serves the needs of all and helps us build a more just, equitable, and enlightened world.