The moral footprint of Generative AI is high, but it is still growing exponentially.
17 Oct 2023, 16:30
The moral footprint of Generative AI is high, but it is still growing exponentially.😱
With that comes potential risks & dangerous implications!⚠️
So, how can you make it safe enough to have even a child play with it?👼
RLHF may have the answer!💯
A thread...🧵
1/9
Same news in other sources
217 Oct 2023, 16:47
But, Safety in AI is as important as its democratization!💯
We at @real_alethea are dedicated to democratizing ownership of AI!
Follow @real_alethea for more on AI and Web3!
RT🔁 to share it with your friends and Bookmark🔖 this for the future.
9/9
But, Safety in AI is as important as its democratization. We at @real_alethea are dedicated to democratizing ownership of AI.
But, Safety in AI is as important as its democratization!💯
We at @real_alethea are dedicated to democratizing ownership of AI!
Follow @real_alethea for more on AI and Web3!
RT🔁 to share it with your friends and Bookmark🔖 this for the future.
9/9
17 Oct 2023, 16:44
Is it really that easy? No!
🎢Scalability: can be time-consuming & expensive
🦓Bias: If human feedback contains biases, the trained AI might also show those biases.
🔊Noise: can be noisy or inconsistent, leading to challenges in training stable AI agents.
8/9
Is it really that easy. Scalability: can be time-consuming & expensive.
Is it really that easy? No!
🎢Scalability: can be time-consuming & expensive
🦓Bias: If human feedback contains biases, the trained AI might also show those biases.
🔊Noise: can be noisy or inconsistent, leading to challenges in training stable AI agents.
8/9