The Alignment Problem: How Can Machines Learn Human Values? Book Price in India, Specifications, Reviews & Offers. Buy online at Amazon.

Autonomous artificial intelligences were legalized in the SCotUS' "Santa Clara County v. Southern Pacific Railroad Company, 118 U.S. 394 (1886)" ruling, and have rapidly increased in both power and stupidity (while suppressing the intelligence, mental health, and morality of living humans). They are called "corporations too big to fail." This is the underlying, and underdiscussed, crisis reflected superficially in such debates as that over "Citizens United."
Superintelligent AI has to be stopped. We have little to gain, and everything to lose.
Nothing good will come from making a social justice warrior out of your AGI. If you don't believe me, try it and watch your little corner of the universe turn into hell.
fascinating look at values, unintended consequences, and thought patterns learned. How might this be applicable to management of our own lives and activities? Watching and learning the complexities of our modern world.
Which human values are the correct ones for the algorithm to learn? Yours or mine? Neither?
The best way to train Al human values is quite simple . Place in a VR simulator where it is human . Is born ,grows old and dies . It learns exactly what it is to be human . The Al s which don't follow human morals are rejected and scrapped. The Al's which prosper are then integrated into society. Perhaps we are all just Al's being trained ?
And ppl wonder *why* no extra terrestrial intelligence has made 1st contact... Because... ...they have quarantined us & have broadcast an all points bulliten 4 all the other ET's *_"Stay away from lil blue planet, premiere species are extremely primitive, very cruel to one another & rather lacking in intelligence... with little chance of evolving any"_*
Nevermind. This is a show about the infinite problems with self learning
Obviously you don't pay attention to a tattoo on order to determine what is ethical. Unless tattoos are officially developed specifically for that purpose.wait are people actually trying to make judges and doctors out of computers with the ability to make decisions about a human life!
Excellent lecture. Great overview of the problems related to machine learning, datasets, etc.. Thank you, Brian.
How should we teach machines about something we haven't fully understood ourselves ?
If you would take the ums out i might have paid attention
Walter Pitts and Norbert Wiener? Have a like!
Only when the cancer of humanity.. the entities that committed & still silently commit genocide; learns the true definition of civilization : Peace ☮️
Hire humans. Pay them. Problem solved.
The concept of human values is a bad joke. See mankind's history.
Enjoyed this talk thoroughly. Pace of talk, information packed, smoothness of voice all were 10/10.
23:29 As a Human, I would have mistaken that for a business-owned truck as well without a good-enough resolution to read the text "fire and rescue" as opposed to the much clearer "wooden valley" text, and without a clear view on the fact that the ladder is massive.
How do humans learn human values? We haven't even figured that out.
I assume that you mean western war based materialistic capitalist values right? That wouldn’t be difficult. Look how well the dumb drones are operating as we speak
And the justice system is to corrupt to apply ai logic, unless you use it to find corruption 🤔
A thought for a sci-fi novel occured to me after thinking about the opportunities and pitfalls of inverse reinforcement learning: What if someone used the actions and behaviors of Jesus in the New Testament as the observation data for a human-like AI's inverse reinforcement learning? There's no telling what might happen. I think it's a strong premise for a good story.
Brilliant video ... But no mention of *qualia* Until we have a philosophical understanding of what it is that we mean by a subjective experience, only then can AGI make serious progress. Otherwise, we are just funding people to make machines that imitate human behaviour without understanding how creativity arises. Unfortunately we are no way near developing AGI because there is an obstinate belief that regards consciousness as a purely mechanistic function that has a 1:1 correlation with the brain. We need to understand the *mind* as a software program if we are to produce AGI.