The politics and philosophy of AI | LSE Event

As artificial intelligence (AI) moves beyond the realm of science fiction, it is already having a profound impact on our economies, societies and politics.

To make sense of the transformative power and disruptive potential of AI, we are joined by thinkers at the cutting edge of the technological frontier for an in-depth discussion of its multifaceted impact.

Speakers:
Professor Geoffrey Hinton
Dr Kate Vredenburgh

Chair:
Dr Robin Archer

#AI #Events #London

To turn on captions, go to the bottom-right of the video player and click the icon. Please note that this feature uses Automatic Speech Recognition (ASR) technology, or machine generated transcription, and is not 100% accurate.

[samuelagboola]

The irony of so many smart people and such amateurish recording. Please fix this. A fascinating and important talk marred by incompetent AV. Given the price of the microphones on display it seems reasonable to expect the LSE invest in an engineer.

[jynxkizs]

While creativity is often seen as problem-solving, what if artistic expression isn't about absolute problems, but about exploiting the human brain's capacity for enjoyment through complex musical structures? This raises the question: is AI simply not sophisticated enough in music theory, or is there an inherent human element to appreciating truly groundbreaking music that goes beyond theory? Are there ethical issues with collecting enough brain data directly, instead of finding patterns in a playlist of songs?

[geaca]

9:22 Dangerous it creates a setting of competition between AI's. Like Dan Hendrycks of CAIS proposes, there should be a neutral, CERN-like international institution for long term Artificial Intelligence development.
9:39 Open sourcing the LLM's is very dangerous too of course.

[musiqtee]

Historically, ANY technology entering our society at scale seem to narrate two stories;

- One, [insert tech] will steal all the jobs...!
- Two, - [insert tech] will liberate us to work less...!

Reality; Three, most individuals end up working (labouring) way more to indirectly cover basic needs via money and debt...!

Work more hours and years, longer education & debt, higher pension age, less pay in nominal terms (ability to buy house, health care, healthy food...). Different - maybe meaningless - work, but never less work.

We call this _growth_, but except for the post WW2 era - GDP is about corporations and governments, not reflecting majority of people or social strata.

Why should Al be different to everything from the printing press (1500's) to steam power (1800's) to networked computers (1990's) ...?

And how will ecology (resources) be transformed into energy (work, not labour) this time - as opposed to earlier in modernity...?

How will ownership of AI models and resulting data balance between the public or private sector and the individual or community…?

I've lost my faith in "progress", as long as everything we create or do must be defined towards capital value. Unless we also ask “for who(m), and for what resources…”?

[emersonpinto]

The oversimplification of a subjective experience (for humans) fits perfectly for a computer model language, however it does not equal to a valid definition of a human subjective experience.
Therefore, how to draw a philosophy of AI without the human subjective experience in perspective? Such a theory of subjectivity devised by a (still) human may not be valid even for an AI modeling of subjective experience.
Most likely, an AI device will learn from St. Tomas Aquinas about the existence of a God as an ontologic fundamental. There is a thousand years of Ancient and Medieval (scholastic) work on ontology to start with.
AI may help on that end, better than academic scholars trying to deny our ontologic fundamentals in a powerpoint presentation.

[loopuleasa]

I just skipped it all, and went to Q&A
seen many hinton talks already