AI Alignment: The Myth of a Superintelligent Savior

There's a buzz in the airwaves about AI, and it's not just the static of existential dread. Talk of AI outpacing human intelligence has reached fever pitch, but let's not kid ourselves. While companies like OpenAI and Anthropic wax poetic about AI's potential to transform the world, there's a solid argument that they're just as interested in lining their pockets.

Tool, Toy, or Just Another Product?

It's not hard to be impressed by the capabilities of AI models like OpenAI's ChatGPT or Anthropic's Claude. They're fun, even useful. But let's take a step back. Are these tools truly poised to replace humans in meaningful ways, or are they simply new toys in the tech playground? The real story isn't about AI taking over but about companies selling products that need to turn a profit.

AI safety, there's a lot of chatter about AI eventually surpassing human reasoning, moving into the world of superintelligence. But who are the 'we' that these AI systems are being aligned with? As it stands, 'we' are mostly profit-driven private companies. OpenAI openly aims to build superintelligence, citing reasons ranging from improving the world to economic growth. Yet, cynically speaking, it's just as likely about making loads of money.

Alignment or Product Development?

Here's where things get murky. The mission of aligning AI with human values gets tangled up with product development. OpenAI and Anthropic are undoubtedly building products, not just theoretical models. These products have liabilities, need to sell, and must capture market share. The press release said AI transformation. The employee survey said otherwise.

For all the talk of alignment, the current research trajectory leans heavily toward making AI sound good rather than be inherently good. Intent alignment, which suggests AI should do what humans want, simplifies a complex reality. Whose intentions are we aligning with, anyway? In a world where financial incentives often overshadow ethical ones, can we really expect these companies to solve the challenge of extinction?

The Real Threat Isn't AI, It's Us

Let's face it: AI's existential risks are hyped in a way that obscures genuine concerns. The story isn't just about powerful algorithms. it's about how we choose to use them. The real risk comes when we start trusting AI with consequential decisions without question. Are we ready to hand over control of our energy grids, utilities, or even our lives to an algorithm because it 'sounds' aligned?

AI's potential for harm doesn't lie in its technical capabilities alone. It's in the choices we make about deploying these systems. The gap between the keynote and the cubicle is enormous. If AI-induced catastrophe is on the horizon, it won't be because the tech got too smart. It'll be because we got too careless, believing in the myth of a superintelligent savior while ignoring the real-world implications.

So, the question isn't how we prevent AI from killing us all. It's how we prevent ourselves from making it so easy.