Top Researcher Li on the Promises and Perils of AI

Fei-Fei Li, Sequoia Professor in Computer Science & Co-Director Human-Centered AI Institute, Stanford University speaks with Bloomberg’s Emily Chang about AI and ethics at Bloomberg Tech in San Francisco.
--------

[architectinth]

To the audio engineer, turn down the gain, and turn up the volume. Basic stuff, basic stuff.

[raphaelamorim]

Fei-Fei Li is a brilliant person. The question about representation in the industry is not valid. It's a derived problem from the distribution of people in the universities. The distributions are quite similar, that's the problem we need to solve first if you want to see any real change in the long run.

[geaca]

I very much disagree with Dr. Fei-Fei Li about not placing importance on possible AI existential threat. When she talks about AI as a humanity collective effort, it should include ensuring safeguards against that.

[mmmch]

i cannot forget her Ted talk about imagenet. 👏

[anre]

She's properly assertive. I like that

[ScotBotAI]

It's amazing how life like she looks.

[nigellawson]

If human beings as a whole have difficulty in behaving ethically. Why should we expect their mind children to behave differently?

[emc]

Asking who she trusts the most or least seems a bit unfair

[Brincat]

build a cad design for making engines with AI creative elvolotions

[DavidK-pu]

Fei Fei Li has her name on over 300 papers, how many of them with her contributions?
Which AI algorithm was created by Feifei Li, if any?
Imagenet is a dataset, like a library, how did the image dataset fundamentally change the AI? Fundamentally?
How much does Feifei Li know about language models?
What are her contributions to AI?

[jamesmhango]

Its comforting to see her confidence. I was watching with girlfriend who also was thrilled. big sisters

[KateShawn-jvwh]

I had initially planned to retire at 62, work part-time, and save money, but the impact of high prices on various goods and services has significantly disrupted my retirement plan. I'm worried about whether those who experienced the 2008 financial crisis had it easier than I currently am. The volatility of the stock market is a concern as my income has decreased, and I fear that I won't be able to contribute as much as before, potentially jeopardizing my retirement savings.

[nigellawson]

Why would a super intelligent AI care about what human beings feel or need? After all, human beings don’t consult chimpanzees before deciding to build a highway through their territory. Therefore, the idea that a super intelligent AI would conforms its actions in order to be aligned with human needs is patently absurd. When it comes down to it, if lower life forms stand in the way of progress, the natural human instinct is to either sweep them aside, eat them, or exploit them in some other way that is of economic advantage to us? Likewise, a race of hyper advanced cyber beings would in all probability adopt the same attitude towards humans, if nature is any guide? Maybe, the best thing we can hope for if this scenario ever comes to pass is that they might take pity on us and decide to keep us as pets? In all probability a super AI might behave towards us in a manner similar to a mega corporation like Apple. On one hand, it might supply humans with all sorts of goodies, while acting in way that is inimical to human needs? Only the future will tell.

[TheManinBlack]

Should we really be asking a start-up founder on what sorts of laws should be passed? Of course they'll be against any kind of regulations.

[kennySg]

Press likes to talk about responsible AI. Hope they can move on to other aspects.

[masterwatch]

Conscious AI "Super AI" will let you know it is real and people will know about it most probably in a way you don't expect. We are not there just yet but there is a technology vector that indicates that conscious AI might be achieved.

[pacanosiu]

The ego will not allow them to do this, it is too strong, only the appropriate procedure will bring the effect, just like with you, now you understand, but your ego still does not allow you to tell the truth about me, do you understand?
Your intelligence is too low and it's not about knowledge, the basis of intelligence is respect + humility + trust

[TaiwanNationStatusQuo]

Why women and people of color matters in this context? We need Responsible Brilliant minds and don't be evil. Gender and ethnicity don't matter.

[HomeEngineer-wmfg]

@13:52 You know the conversation was pretty good up until "The Message" mantra of "....women and people of color...." and then

[paideia-eu]

Humanities have not demonstrated the ability to achieve lifelong happiness. They are not able to overcome their inner and hidden psychological sicknesses, weaknesses, and self-contradiction because they are not fully conscious of these. People have built AI robots strong enough to act out human self-contradiction and self-destruction. No long-term happiness or longevity will exist if people don't wake up from their unconscious passiveness.