#14 Monday Note: I am so worried about AI bias, but not in the way you think.

When we talk about AI bias, we often focus on the flaws and risks associated with these new technologies. It’s human nature to be skeptical of new things and to find comfort in pointing out their imperfections. How often have you heard, “Look, it can’t even create PowerPoints?” Or “It’s just another technology,” and “Show me the business case.” This, my friends, is bias par excellence.

The reality, especially with large language models (LLMs), is quite different. We barely scratch the surface of understanding their capabilities and the far-reaching implications they already have. As Ethan Mollick points out in his blog about LLMs in science, LLMs, when used skillfully, can even create new mathematical knowledge on their own or enable new forms in social sciences (link in the comments).

The challenge isn’t just in improving these technologies but in overcoming our own biases against them. It’s crucial to recognize the extraordinary potential that lies within LLMs and other AI advancements. Instead of dismissing them for their current limitations, we should be exploring and pushing the boundaries of what they can achieve. The future of AI holds tremendous promise, and it’s up to us to harness it responsibly and creatively.

#AI #Bias #Human #Understanding #Science #LLM