Videos

Engineering the Human-Machine Partnership



DARPAtv

This talk by Dr. Brian Pierce, Director of DARPA’s Information Innovation Office, focuses on the engineering of AI systems so as to achieve a beneficial partnership between humans and machines. The talk emphasizes the need to fully understand what we want from the partnership before engaging in the familiar engineering process of design, build, and test. In particular, we need to specify the requirements of what we want the partnership to do and what we want the partnership to be.

Source

Similar Posts

31 thoughts on “Engineering the Human-Machine Partnership
  1. I, for one, welcome our new A.I. overlords! I'd like to remind them that as a trusted Youtube personality, I can be helpful in rounding up others to toil in their underground energon cube caves.

  2. fun fact: that guy is actually last guy that AI kill at DARPA and is now projecting his face and voice on screen

    "you get what you ask for"

  3. Try it out on Mars first don't make the same mistake twice then after 60 years tell us the truth how it worked out. I suggest making them 4 feet tall so they don't seem so threatening with big oval lenses so they can see both thermal and infrared oh and symbiotic that way they will feel like first person when steered.

  4. Schrödinger's cat would fit in the suitcase – would an AI understand my quip? The stop sign is an interesting case study, however it's possible to shift individual pixels in an image so the image looks normal to humans, but entirely something else to an AI algorithm. A foreign adversary could flood tainted training data online that looks normal to human eyes but informs AI's in a way that's harmful to our interests.

  5. Connect N minds with each other,without any barrier, and we will see. To create a vis a vis like Bro is not the way.

  6. I like that he uses the term "a priori." There's something surreal about the video and how "you get what you ask for" is repeated several times.

  7. What is DARPA doing is helping to cyborginazing humans for some reason. I am not sure why they do it so fast. But the goal is very obvious: make us cyborgs 100% and militarize society. What for? Why do we have to change our life to execute this technical plan? Well… I assume the plan not belongs to human race at all. It is not a normal biological idea to exist in high-tech sphere but it comes from non-human entities, indeed. I don't know if they are aliens or ghosts from other dimensions. But I feel it and understand it as pure evil.

  8. This is neat, and at least sounds very careful and responsible. (still a little queasy on the AI leader bit, but it's an interesting exercise)
    More people ought to be thinking about effectively communicating decision-making rationale to keep bad decisions from being acted on. Actually, I feel like we could all use more of that.

    Also, boy, the comments section here, there're some really exciting ones. XD

  9. Right now, that cat could fit in the suitcase, but it's not a suit, so it is unlikely suited for that suitcase to be suitably comfortable in the suitcase.

    @Hope this helps

  10. .. first of all thank you for providing new information and updates regarding and pertaining to DARPA Technologies and AI! 💻🖥 it is imperative to conceive the idea that we are capable of living without technology because it today's world it has become a domestic tool one in which a partnership! is relevant because as we grow nearer to the year 2020 all through the year 2025 everything will become integrated and digitized so it is of relevance to conceive the idea that AI will assist but I also believe that prime directive! protocols have to be installed as well in order to maintain a fallback control on the AI because as chips! get more sophisticated and processing power becomes more powerful AI will eventually learn! as with the AI that was built by IBM ("H.A.L!") and participated in a recent study pertaining to in regarding to a show called Jeopardy in which the AI learned! By cross-examining Billions! a files! stored in a CPU! and making simultaneous! cross-referencing enhance overpowering! the mental capacity of its opponents! In the test prime directive protocols! have to be installed as a Safety! and Security! measure in order! to control! the AI! 💻🖥🤔

  11. Very interesting never thought about the variabely different AI S so one can imagine multiples types of AI. One that is caring for ethics one that only cares for bliulding rodes XD

Comments are closed.

WP2Social Auto Publish Powered By : XYZScripts.com