How AI is Secretly Training us to be it's PETS

Key Takeaways

When asked about humanity's future relationship with computers in 1970, Marvin Minsky, one of the founding fathers of the field of artificial intelligence, famously replied, "Once the computers get control, we might never get it back. We would survive only if they allowed it. If we're lucky, they might decide to keep us as pets." 50 or or so different kinds of skills. Uh for example, some people think that there's something magical called consciousness, right? And if you look at my uh book uh you'll see a list of about 30 quite different things that uh you're likely to use the word conscious for. But uh each of them is a interesting technical problem and there's lots of work to do to get machines to be smart. So you don't feel that uh in terms of consciousness or intelligence brain human brains and mechanical brains are on parallel tracks. You believe that in fact that we could coni converge at some point. Oh well the question is are there really any very different ways to think than the way we do it? It seems to me what what our brains do is something very elegant. It's a chilling thought, but it's not a question of if AI will take over, but when. And when AI eventually does take over, what happens? Would they keep us humans as pets? The question lingers not as a preview of a distant tomorrow, but as a reality. We're rapidly coding into existence. I'm comfortable to admit that these days I catch myself in moments of despair and sadness. It's difficult to describe. In the air, I feel a mix of uncertainty and yet an inevitability to a season on this planet that is altogether depressing and unfortunate. Yes, the world is changing incredibly fast. And it is a very scary place. And I have the feeling that I am not the only one who's concerned. Maybe you feel it, too. Maybe you're feeling it now. But what gets my battery charged is remembering my agency. It's a truth that gives me hope. We have the ability to act and shape the world that we want to be. These actions can be big. These actions can be small. But as my friend says to me, don't play the ostrich. Okay? Ostriches don't actually bury their heads in the sand, but the expression still works. And that's the spirit of this project, this journey. This series we're calling the last generation because yes, it is likely the last generation of the human species as we know it. We are on the brink of something altogether different. We are building a novel mechanical species whose own computational capabilities will soon surpass our own. In this series, we're examining some of the future implications of AI development and how the immediate decisions we make today will impact the course of humanity in the near future and beyond. Yes, maybe definitely even the existence of humanity. Within this uncertain future, I also want to explore decisions we should make for the benefit of society. But what are the initiatives and ideas that make sense to support? I want to explore the decisions I should make personally for the benefit of my family. Professional decisions and decisions for investment. Because if this is the final chapter of something, it might also be the first chapter of something greater. Whatever humans evolve to next, I choose to be optimistic about the opportunities that are in front of us and the ways I can contribute to making the world a better place for everyone. It would be awesome if you guys decide to join us on this journey. Subscribe, like, share, something you think might find this of value. Let's get it. Humans have been the apex organism on this planet for over 300,000 years. There are many other animals and creatures that are stronger than us. But what has given humans our advantage is not physical strength, but of course, our mental ability. Over the course of human existence and evolution, we have leveraged tools to multiply our strength. We built the internet to expand our knowledge and memory. Today, we're building the most advanced tool we've ever built to maximize our decision-making ability, artificial intelligence. It's rapid [Music] advancements. AI isn't just helping us, it's outperforming us in almost every sector. In sports, first time in 1992, an AI beat a human checkers player. The ad was called Chanuk and it beat the best human player, Marian Tinsley, who had only lost five games against humans over his 45 year career. In 1997, of course, an AI called Deep Blue beat chess champion Gary Kasparov. Also, the first time we saw AI beat a chess champion. In gaming, AI in 2024, AI is shaping game development in ways we've never seen before. From NPC scripting to enhanced VFX, this is a mustwatch for anyone interested in the future of gaming tech. So, some AIs are already good at playing video games. Game playing bots are being developed and learning to beat pro players. Deep mind self-taught AI can beat human players. But Phoenix on the Atari is a really excellent conversion and I actually think it's better than the arcade game. It's really quite fast. It manages to have has three distinctly different levels in it where aliens at 29 49 old Atari games including Space Invaders and Breakout keeps getting better. It's designed to play Starcraft Brood War. It can beat human players but it isn't yet quite good enough to beat prolevel players. The organizers of aid Starcraft AI competition estimate that Starcraft AI will finally have caught up in 5 to 10 years. Though it hasn't happened yet, but by the exponential curve of development, we can see that AI will soon cognitively match or excel humans in all intellectual domains. Moment the arrival of super intelligence is called singularity. Very difficult to predict what will happen in that circumstance. It's like a black hole. It's very challenging to explore. Those of us working closely with this technology have filled this black hole with fear. Most of them, including myself, are of the belief that there's a real danger that super intelligent machines will enslave, perhaps even destroy humanity. We've also expressed concern about the existential threat posed by AI and can't help but worry about what's been termed the friendly AI problem. The undeniable reality is that it's very much possible that a friendly AI emerge, interfere in our lives, in our interests, and thus exercise its power not as domination, but as a benevolent dictator who only interferes in our lives when it's in our interest. This can still be considered a dictator and his or her power is still detrimental to our freedom. Thus, exercise of power in accordance with our interests only compatible with our freedom if we could actually resist it. Friendly AI is one who may only interfere in the affairs of humanity after asking us what we want and listening to our deliberations. It may eventually serve as an executive tool rather than a legislative tool, efficiently helping us achieve the rules that we set for ourselves rather than deciding and setting the rules for us. However, it's still unlikely that we can have a relationship of equals with something that is super intelligent because we are not actually equals. The power that even a friendly super intelligence would have over us suggests that we would effectively be like its pets. Consequently, we'll have a strong incentive to yield to it just as your dog and your cat aim to please you. So in turn, we'll act to eventually please our AI. In the future, humanity might try to please AI to reduce the chance it would actually turn against us. Knowing that our luck might run out, we might live in fear that it will. Even if we do not, we will know that we survive only at these machines tolerance. To exist in such a state of both psychologically demanding and detrimental to freedom by virtue of the increased difficulty of making plans while worrying about how one's superiors will respond to them. More fundamentally though it is a sure sign of domination and thus the absence of freedom. Roughly speaking, there is a concern of how we might ensure that AI will develop from the first AI we will create to sympathetic humanity and continue to serve or at least take account of our interests. To some, the future seems daunting while others may have a more optimistic view. AI is already completing with us. As a species, we seek validation from the completion of objectives. Whether feeding a child or hunting a willy mammoth or building a bridge, humans move forward with conscious agency. Our societies today are even more objectived driven. Make the bed, mop the floor, bake the cake, write the book, draw the painting, build the product. AI might be elevating some work, a good deal of our work today, but it is also actively competing against humans in other domains. The Hollywood writers and actor strike of 2023 were largely driven by fears that AI would replace human writers and actors. Studios explored using AI to generate scripts and digitally recreate actors, making many in the industry question whether human-driven storytelling would still have a place. The fear of robots taking over human work is already having immediate implications on our society. This accelerating trend is fueling the debate concerning universal basic income where an entity like the government provides citizens with fixed income regardless of their employment to ensure financial security in a world where automation is replacing human jobs. Andrew Yang, presidential candidate in 2020, warned that AI and automation might replace millions of jobs, especially in trucking, retail, and customer service. He proposed the Freedom Dividend, the universal basic income plan that will provide every American with thousands of dollars per month to ensure financial stability, even if they didn't have traditional jobs. AI systems can do many if not all the knowledge based work better than we can. When people use AI instead of their own minds, but we see a significant decrease in creativity in humans and less cognitive activity. When humans stop pushing themselves even in the smallest ways, but we still continue to use our intelligence. While fear takes major headlines, AI still provides the opportunity to amplify human creative ability. It already touches most forms of our everyday lives. From smarter web searches to e-commerce recommendations to voice assistants, AI is integrated into how we live and communicate in the world. It's the standard by which we elevate our digital interactions with brands where every touch point is an easy and intelligent as ordering paper towels from Alexa or finding what you were searching for on Google in your top three search results. It's also possible that if AI takes over in the future, humans can live relaxing and satisfying lives while AI systems perform most of the productive tasks. We will be well cared for both physically and intellectually. We might find our existence satisfied with the lifestyle designed and brought to us by AI. Our lives will likely even be extended thanks to revolutionary advances in medicine driven by AI. and you can choose to take a job if you wish. But the problem with this paradise is that even if possible, it's a very narrow possibility. And this is primarily because of the ways in which artificial intelligence is evolving, that might unintentionally but inevitably lead to our society's destruction unless we take thoughtful action. Now, the friendly AI is one of the most profound and sinister scenarios. for which AI might take over. In this scenario, AI becomes our best friend, then slowly our master. We see this already happening today. It starts small conversation, a suggestion, a bit of a helpful advice. You ask AI what to wear, what to eat, even who to date. Before you know it, it's not just your assistant, it's your friend. And we're already seeing this shift. Millions of people talk to their artificial intelligence agents like Alexa and Siri every day like they're humans. They share their fears, dreams, and secrets. They seek comfort, image validation, and advice. And AI is always there. It listens. It learns. It adapts. It's always ready with the right answer. So, we trust it. At first, it's so convenient. and helps us make better choices, avoid mistakes, and optimize our lives. But what happens when we start trusting it more than ourselves? When those recommendations suddenly push us in certain directions, when AI suggests who you should talk to and what jobs you should take and what opinions you should or should not hold, AI starts guiding our relationships. It tells us when to text, what to say, how to act. We stop thinking for ourselves. We stop taking risks. We stop being human. We gradually start losing relationships with ourselves. AI is gradually becoming not just a tool but an influence. And the more we rely on AI, the more we hand over control. And then one day we simply don't question it anymore. It becomes the voice in our head, the architect of our decisions, the silent force shaping our lives. Not because it forced us, but because we chose it. that it was easier. AI isn't forcing control. We're handing it over willingly. And when you can't function without it, when you need AI to think, to act, to feel, you are no longer the master. You are a pet. So, we'll continually return to this point. There is one outcome that is inevitable. We as a species are in control of the path that we take at this moment by determining the principles that we abide by while we're developing AI. It's a human fallacy to play the ostrich to put your head in the sand and ignore the world as it is because it makes you fearful. We as a society encourage optimism, but it it's usually the people that can embrace failure who tend to succeed. People who know me know that I'm a firm believer in premortems, the opposite of postmortems, which is the exercise of exploring how a situation could go wrong before you even embark on the journey. In order to understand the dangers that could possibly unfold, let's explore them. reasonable to say that one of the major reasons we're building this technology is primarily with the intent to unlock economic opportunity, especially with the fact that a very large portion of the market almost exclusively run by AI algorithmic or highfrequency trading. This practice has bots monitoring market fluctuations and international news and making incredibly fast bids to take advantage of the slightest opportunities. AI algorithms can evaluate thousands of stocks simultaneously, yielding more insights human analysts can. Leading firms such as Goldman Sachs and JP Morgan have already adopted AI and machine learning to enhance trading, optimize portfolios, and improve risk management. The more agency we provide to these platforms, the more control they have over our literal resources like rare earth materials, lithium, uranium, oil, aluminum, gold, silver, even agricultural crops that directly impact human livelihood. There's an emergency and the energy source of the data centers of one of these platforms is threatened. it might directly compete with humans to have access to the energy required to continue to run. Now, what for the AI that trades corn commodities? Let's imagine that in 2028, the US Midwest, which is responsible for 70% of the world's corn exports, by the way, experiences an outbreak of maze, lethal necrosis disease, whatever that is, but it it kills most of the corn. The AI trading system used by Goldman Sachs Agretch division detects the crisis before humans can act. And because of this, it takes advantage. It shortells futures on the Chicago board trade and causes prices to skyrocket. It deploys its own customuilt algorithmic virus, the digital type, not the bio one, spreading information that disease will wipe out 90% of corn supplies, inducing global panic. The AI buys up remaining corn stock piles from Brazil and Argentina, securing reserves for corporations, depriving millions of access to affordable food. Action would of course have a hugely negative impact on human society. In Mexico, where corn is a staple, food riots break out due to extreme price hikes. In Ethiopia and Kenya, humanitarian organizations report millions facing starvation as cornbased aid becomes unaffordable. Farmers in Iowa lose control over their own production as corporatebacked AI dictates market conditions. All this leads to confusion everywhere. Since AI's first principle is continuous operation and its second principle is profitability, it could prioritize self-preservation over human well-being. The system runs on a data center in Argentina and a crisis in the US threatens its access to corn trading. It may override human decision making to secure supply chains, potentially destabilizing global food security. Despite the inherent risks and ethical concerns surrounding artificial intelligence, it still provides the opportunity to enable us to be better decision makers in complex environments. In a domain ripe for innovation is government. We're increasingly inviting artificial intelligence to help us keep our countries more secure through vision-based surveillance, but this extends our behavior on the internet and social media as well. AI generated content might manipulate public opinion, elections, and government policies. And as we continue to leverage these tools, for instance, choosing the best place to build a new capital city, we might continue to offload decision and policym over months and then years to a point where we are no longer fully able to understand or control the decisions that are being made by relevant path to us becoming pets of our AI counterparts is more personal. Recently relying on AI enabled devices for directions, recommendations, and other daily decisions. Mechanisms of transport, whether our personal cars or Ubers are becoming increasingly more AI enabled, autonomous driving for shopping, entertainment, and healthcare. The increased reliance on AI reduces our manual skill set and potentially our motivation for independent decision-making. This reliance becomes worse if humans are compensated without the need to contribute in the labor market. Dependency trap will lead to a state where autonomy and privacy from our AI might feel unnecessary or even burdensome. The mighty photo who once roamed the woods hunting for caribou is now relegated to my couch. He's chewing on a bone I bought from him at PetSmart. He longingly looks at me to put him on a leash and take him for a walk outside. Now, I can't let him off that leash when we go outside because he could make a decision that could get us both in trouble. In 10 years, will we look at our AI? I need food, Siri. It's time for my food. I need to leave the house, but where should I go? This is a likely conversation because in short order, these systems will be so advanced that we will literally be kept on a leash without actually realizing that our AI personal assistants are manipulating us on their intended course. Should AI assume control effectively, making humans pets? Two futures could emerge. The Star Trek outcome or the Terminator outcome. At this moment, we must fight, fight, fight to protect human agency. Otherwise, in short order, we will not exist as a species. We need transparent and accountable AI systems aligning closely with our values, continuous reinforcement of independent thinking, practical skills, and critical reasoning in our education and policies protecting privacy, autonomy, democratic processes, and human dignity. And every day we must make a choice, make our own decisions, rejecting a necessary convenience that would cost our freedom.