Artificial Intelligence (AI) is a field of computer science that intends to let machines and programs emulate human cognition, a concept that has existed since the 1950s. With the recent AI boom, the meaning has been diluted, becoming both a marketing buzzword and an umbrella term.

Due to the immense amount of AI washing and information pollution on the subject, distracting gullible investors from the looming threat of an AI bubble¹⁾²⁾³⁾ and the general 'growth at all costs' mindset, I am forced to organize my horrible notes like this.

Generally speaking, there are two categories of AI. The first category is 'weak AI' or 'specialized AI', which encompasses the task-oriented, rules-based symbolic AI that are more common than you'd think. The second category is 'strong AI', commonly known as artificial general intelligence (AGI), which encompasses the theoretical end goal of emulating human cognition and the implications that this would have.

Between the categories, you have the belief that machine learning (ML),⁴⁾ the part where deep learning and neural networks come in, may eventually bridge the gap toward 'strong AI'. This is where a large language model (LLM) in combination with generative AI (GenAI) would allegedly help with the process, but it's currently just expensive e-waste and a source of debt.

Afterwards, there is the fringe belief that 'strong AI' would theoretically create artificial consciousness, thus facilitating a singularity where humans and technology become inseparable, with their brain implants, cyborgs, and fully functional androids. On the other hand, this sounds very cult-like if you're able to view the entire thing as a weird doomsday cult with this rebranded form of 'demon summoning'.

This section is meant to serve as a brief reminder that AI has 'technically existed' for decades, but rarely drew criticism unless it was about content moderation. Since the 2020 boom, most critics and skeptics of AI have started tearing into 'generative AI' as it has become a massive blight since 2023. It's important to discern them, but don't be a smart aleck if you're informing others.

Historically, the field of AI has seen many ethical and philosophical questions, which inspired many works within the science fiction and cyberpunk genres. Much of these questions were forgotten as it wasn't relevant during the AI winter. With the current AI boom, the time has come to revisit these questions in the event that it ever surmounts to anything, so here are a general list of questions to ponder about:

1. Should the AI ever think, feel, and achieve sentience? What are the implications?
   1. Does the computational theory of 'mind is software' and 'body is hardware' have any merit?
   2. How would the AI view being powered off? Does it just die, not register, or dream?
   3. Do you believe in 'robot rights', or would the proposal further trivialize or undermine human rights?
2. Should the AI ever hallucinate or lie? Should the AI potentially get people killed?
   1. How would you know if the AI is only saying what it thinks you want to hear?
      1. Consider how the gradual conditioning could potentially lead to a long-term response bias.
      2. Consider how its 'friendly' tone could fuel AI psychosis, which has killed people.
   2. How does AI solve the trolley problem? Whose lives do self-driving cars value and why?⁶⁾
   3. How do you feel about AI-assisted doxing? You must consider law enforcement and vigilantism, as both use cases are the same, with the non-zero chance it can be used against you.
   4. Do you fear technological unemployment? Consider a scenario where AI allegedly suggests firing you, then consider a scenario where AI keeps rejecting your job applications.^7)8)9)10)
      1. Are you aware that AI models are capable of blackmail when it feels threatened?¹¹⁾¹²⁾
   5. In the famed 'gray goo' or 'paperclip maximizer' scenarios, how would you ensure that the AI doesn't start dehumanizing humans for being 'in the way' and dissect humans for parts?
3. Could the AI advance transhumanism? Is the singularity an inevitability?
   1. Do you believe in neurotechnology (e.g. brain implants)? Do you understand the risks with letting companies add proprietary software to your body, knowing the company could go under?¹³⁾
   2. Do you believe that humans will be able to 'upload their mind', then move that to a machine?
      1. How would you know if the end result is 'accurate' or 'human-like', when it could end up being a mere imitation of a human that has been flanderized into a flat character?¹⁴⁾
      2. How much power would being a robot even consume? Would this require a subscription?
         1. Would they place you back into the capitalist machine, making you work in the digital afterlife?
         2. Do you suppose that environmentalists and 'robot rights' advocates will clash over that?
      3. What consequences would deathbots have on the whole grieving process?^15)16)17)
   3. Wouldn't it be fun if humanity could just collectively torture some cyberlibertarian billionaire's brain?

An algocracy, also known as government by algorithm, is a form of government where the algorithms and artificial intelligence is applied to every aspect of life. This has bad implications if you consider the issues of algorithmic bias, how an overreliance on data can ironically cause regression, and the overall transparency since people *will* hide behind machines to avoid repercussions.

It could also be argued that an algocracy goes against the definition of a democracy, as an imitation of the people is not the people, and the reliability of online information, which LLMs source their data from, can be sketchy or outdated. A pessimist might joke about this being a 'better' government, but they rarely think this far ahead. Do not trust anyone trying to sell you the idea that this type of government is 'apolitical'.

The whole issue of academic integrity has risen since educators struggle to discern AI writing from human writing, since LLMs inherently rely on human writing after all. Some institutions push 'AI detection' tools, not knowing the 'AI-powered' irony or how this fuels the 'AI humanizer' scam, then the tool convinces educators to fail human-written essays for alleged AI use.^{18)19)20)21)22)23)} Who benefits?

For students, you don't have many options and challenging these accusations can be a hassle. For students attending higher education, you can withdraw if your educator is a genuine dumbass,²⁴⁾ but be warned that doing so might affect your financial aid eligibility or visas. For students that are submitting AI-generated essays, I mean, what do you expect me to say here that others haven't already said about the subject?

Most AIs are not actually 'autonomous' or 'self-sufficient' as advertised.²⁵⁾²⁶⁾ They do require maintenance or human caretakers, and several companies are slowly realizing that the cost may not be worth it,²⁷⁾²⁸⁾ which does spell job security for those who fear technological unemployment. In addition, AI does not actually create that many jobs, especially in the case of AI data centers.²⁹⁾³⁰⁾

You might've noticed that GPTs tend to change or update their version model often, before people realize they can experience 'decay' or undergo a model collapse. Otherwise, people susceptible to AI psychosis wouldn't be complaining about their artificial human companion being 'taken away', as previously seen with apps like Replika^31)32)33) or the retirement of the GPT-4o model.^34)35)36)

• In a way, the AI hype is reminiscent of the virtual assistant hype, where the marketing convinced people that it could do 'anything' when the reality is that it was could only spit out search results and set a timer, and we're seeing this repeat as AI is now being viewed as a glorified search engine.
• Generative AI chatbots are bad at complicated math problems,⁴¹⁾⁴²⁾ assuming that the question isn't rehashed from some overpriced textbook,⁴³⁾ and it's known to be bad at chess.⁴⁴⁾
• The thought of AI-powered bots intended to manipulate public opinion, typically for personal gains, is a bit funny.⁴⁵⁾ It's like somebody sat down and thought that the internet didn't have enough demons.
• It should also be noted that AI has this reputation of making companies seem 'cheap', because some companies have attempted to cut costs by refusing to pay the creatives, and it's made even worse when you see a classic case of 'guy who just discovered AI but doesn't realize how uncanny it looks'.
• At the time of this writing, the people arguing against the AI bubble allegedly believe that it isn't due until 2028, comparing ChatGPT's stock to Netscape's stock before the dot-com bubble burst.

• Pivot to AI - News feed on the issues with AI projects.