Nada School Saturday Class 2025

2025-05-24

https://gyazo.com/43c952a7f0f8fafb60a2a9b4e671a7ae

https://gyazo.com/2a3c6605e8ecdc90b564dcf8daaea242

I'll just reprint my part.

https://gyazo.com/5ca2f96d27bcc44530548a7baf8a55c7

https://gyazo.com/ca71021c4f240dac8e3b1ba22660c203

https://gyazo.com/61688742f2bbdb577df4d3320577357e

https://gyazo.com/5b0306c62c00c02873809829186e4ee1

https://gyazo.com/558069a4c9bb07e054274955c61479c5

https://gyazo.com/7f49dd2fce2f19a7460fed35aecdd434

https://gyazo.com/157295ad87669435b6eee0e7ef268499

https://gyazo.com/3ceb5c9d20f1ef364bdd07c797b6eb0e

https://gyazo.com/02df5c5ebaab8415a215c8900b2aa276

https://gyazo.com/e472a026b531726dfc6d7be4ffb238a7

https://gyazo.com/e696d2da41bca9d63dadccf31e508bda

https://gyazo.com/2138949daacfd7b08383daecce2b1656

https://gyazo.com/802b6ebe62f16fa089bc6f92002eec9a

https://gyazo.com/a6980af8ca3a75b9fec33e67b16a4f4e

https://gyazo.com/c2feb3a23fccf9c6b81a909de30c3008

https://gyazo.com/9a6ad6afaa929a7e4cd2329f4e60f9aa

https://gyazo.com/ccc9759dad53e71a9a69caf81f3863bf

https://gyazo.com/2a213a0a7e0c1c8615b2ff3a0cf36f9c

https://gyazo.com/e29e794bd7b4b81bdcfe8b86c8d36fc6

question

Isn't there a possibility that the AI could skew it when it collects the opinions of many people and puts them together?

Of course there is that.

Related question: is there a bias in the LLM?

Diverse biases by LLMs.

But is there no bias if it is put together by a human being? Of course, humans have biases.

So this is not an "AI or human" issue.

Issues that should be resolved by creating mechanisms such as making the process open and verifiable by a third party

Related question: will AI be a black box?

Research is also underway to examine the internal firing of the model during task execution. This is like sticking electrodes into the brain if you were to use a human analogy.

Stabbing electrodes into public officials is ethically difficult, but disclosing the internal state of AI in public service will be possible in the future.

If you think being a black box is a problem, you'd rather use AI instead of humans.

Not all people can think and speak out about politics, what do we do with those who don't?

The current lack of opportunities for people to have their opinions heard and to feel good about expressing them is the reason why the ability to think and express opinions is not developed.

In Taiwan, systems such as JOIN created opportunities for students to voice their opinions from the student stage.

Similar clusters were seen in [Shin Tokyo 2050 Broad Listening

Today's major social networking sites are marked by people with radical opinions.

Most people don't want to argue with radicals.

Influence the international community?

Of course, the elimination of language barriers is significant

Q: If the generated AI creates a manifesto based on "what opinion will be received by citizens", won't all parties be the same?

When they gather in large clusters, differences disappear and supporters are more likely to be in flux, which the existing parties do not like.

When you visualize it, there are clusters that are not represented by a political party even at this point in time.

Even a small cluster can become a political party with 1%.

The assumption that all parties should go to large clusters is incorrect because it is a red ocean to go for a majority.

In the long run, it's unclear whether party politics will continue to feel the same way as it does now in the first place.

The balance of advantages and disadvantages of the party line is affected by the balance of advantages and disadvantages of the party line.

It may not be stable if the support base changes from one place to another.

I'm not sure where you're coming from with the assumption that it won't change, that it will change from one thing to another.

---