English subtitles

← How AI can enhance our memory, work and social lives

Get Embed Code
35 Languages

Showing Revision 6 created 08/16/2017 by Brian Greene.

  1. I'm here to offer you
    a new way to think about my field,
  2. artificial intelligence.
  3. I think the purpose of AI
  4. is to empower humans
    with machine intelligence.
  5. And as machines get smarter,
  6. we get smarter.
  7. I call this "humanistic AI" --
  8. artificial intelligence
    designed to meet human needs
  9. by collaborating and augmenting people.
  10. Now, today I'm happy to see
  11. that the idea of an intelligent assistant
  12. is mainstream.
  13. It's the well-accepted metaphor
    for the interface between humans and AI.
  14. And the one I helped create
    is called Siri.
  15. You know Siri.

  16. Siri is the thing that knows your intent
  17. and helps you do it for you,
  18. helps you get things done.
  19. But what you might not know
    is that we designed Siri
  20. as humanistic AI,
  21. to augment people
    with a conversational interface
  22. that made it possible for them
    to use mobile computing,
  23. regardless of who they were
    and their abilities.
  24. Now for most of us,

  25. the impact of this technology
  26. is to make things
    a little bit easier to use.
  27. But for my friend Daniel,
  28. the impact of the AI
    in these systems is a life changer.
  29. You see, Daniel is a really social guy,
  30. and he's blind and quadriplegic,
  31. which makes it hard to use those devices
    that we all take for granted.
  32. The last time I was at his house,
    his brother said,
  33. "Hang on a second, Daniel's not ready.
  34. He's on the phone with a woman
    he met online."
  35. I'm like, "That's cool, how'd he do it?"
  36. Well, Daniel uses Siri
    to manage his own social life --
  37. his email, text and phone --
  38. without depending on his caregivers.
  39. This is kind of interesting, right?
  40. The irony here is great.
  41. Here's the man whose relationship with AI
  42. helps him have relationships
    with genuine human beings.
  43. And this is humanistic AI.
  44. Another example with
    life-changing consequences

  45. is diagnosing cancer.
  46. When a doctor suspects cancer,
  47. they take a sample
    and send it to a pathologist,
  48. who looks at it under a microscope.
  49. Now, pathologists look at
    hundreds of slides
  50. and millions of cells every day.
  51. So to support this task,
  52. some researchers made an AI classifier.
  53. Now, the classifier says,
    "Is this cancer or is this not cancer?"
  54. looking at the pictures.
  55. The classifier was pretty good,
  56. but not as good as the person,
  57. who got it right most of the time.
  58. But when they combine the ability
    of the machine and the human together,

  59. accuracy went to 99.5 percent.
  60. Adding that AI to a partnership
    eliminated 85 percent of the errors
  61. that the human pathologist
    would have made working alone.
  62. That's a lot of cancer
    that would have otherwise gone untreated.
  63. Now, for the curious, it turns out
  64. that the human was better
    at rejecting false positives,
  65. and the machine was better
    at recognizing those hard-to-spot cases.
  66. But the lesson here isn't about
    which agent is better
  67. at this image-classification task.
  68. Those things are changing every day.
  69. The lesson here
  70. is that by combining the abilities
    of the human and machine,
  71. it created a partnership
    that had superhuman performance.
  72. And that is humanistic AI.
  73. Now let's look at another example

  74. with turbocharging performance.
  75. This is design.
  76. Now, let's say you're an engineer.
  77. You want to design
    a new frame for a drone.
  78. You get out your favorite
    software tools, CAD tools,
  79. and you enter the form and the materials,
    and then you analyze performance.
  80. That gives you one design.
  81. If you give those same tools to an AI,
  82. it can generate thousands of designs.
  83. This video by Autodesk is amazing.

  84. This is real stuff.
  85. So this transforms how we do design.
  86. The human engineer now
  87. says what the design should achieve,
  88. and the machine says,
  89. "Here's the possibilities."
  90. Now in her job, the engineer's job
  91. is to pick the one that best meets
    the goals of the design,
  92. which she knows as a human
    better than anyone else,
  93. using human judgment and expertise.
  94. In this case, the winning form
  95. looks kind of like something
    nature would have designed,
  96. minus a few million years of evolution
  97. and all that unnecessary fur.
  98. Now let's see where this idea
    of humanistic AI might lead us

  99. if we follow it into
    the speculative beyond.
  100. What's a kind of augmentation
    that we would all like to have?
  101. Well, how about cognitive enhancement?
  102. Instead of asking,
  103. "How smart can we make our machines?"
  104. let's ask
  105. "How smart can our machines make us?"
  106. I mean, take memory for example.
  107. Memory is the foundation
    of human intelligence.
  108. But human memory is famously flawed.
  109. We're great at telling stories,
  110. but not getting the details right.
  111. And our memories -- they decay over time.
  112. I mean, like, where did the '60s go,
    and can I go there, too?
  113. (Laughter)

  114. But what if you could have a memory
    that was as good as computer memory,

  115. and was about your life?
  116. What if you could remember
    every person you ever met,
  117. how to pronounce their name,
  118. their family details,
    their favorite sports,
  119. the last conversation you had with them?
  120. If you had this memory all your life,
  121. you could have the AI look
    at all the interactions
  122. you had with people over time
  123. and help you reflect on the long arc
    of your relationships.
  124. What if you could have the AI read
    everything you've ever read
  125. and listen to every song
    you've ever heard?
  126. From the tiniest clue,
    it could help you retrieve
  127. anything you've ever seen or heard before.
  128. Imagine what that would do
    for the ability to make new connections
  129. and form new ideas.
  130. And what about our bodies?

  131. What if we could remember the consequences
    of every food we eat,
  132. every pill we take,
  133. every all-nighter we pull?
  134. We could do our own science
    on our own data
  135. about what makes us feel
    good and stay healthy.
  136. And imagine how this could revolutionize
  137. the way we manage
    allergies and chronic disease.
  138. I believe that AI will make
    personal memory enhancement a reality.

  139. I can't say when or what
    form factors are involved,
  140. but I think it's inevitable,
  141. because the very things
    that make AI successful today --
  142. the availability of comprehensive data
  143. and the ability for machines
    to make sense of that data --
  144. can be applied to the data of our lives.
  145. And those data are here today,
    available for all of us,
  146. because we lead digitally mediated lives,
  147. in mobile and online.
  148. In my view, a personal memory
    is a private memory.

  149. We get to choose what is and is not
    recalled and retained.
  150. It's absolutely essential
    that this be kept very secure.
  151. Now for most of us,

  152. the impact of augmented personal memory
  153. will be a more improved mental gain,
  154. maybe, hopefully, a bit more social grace.
  155. But for the millions who suffer
    from Alzheimer's and dementia,
  156. the difference that augmented
    memory could make
  157. is a difference
    between a life of isolation
  158. and a life of dignity and connection.
  159. We are in the middle of a renaissance
    in artificial intelligence right now.

  160. I mean, in just the past few years,
  161. we're beginning to see
    solutions to AI problems
  162. that we have struggled
    with literally for decades:
  163. speech understanding, text understanding,
  164. image understanding.
  165. We have a choice in how we use
    this powerful technology.
  166. We can choose to use AI
    to automate and compete with us,
  167. or we can use AI to augment
    and collaborate with us,
  168. to overcome our cognitive limitations
  169. and to help us do what we want to do,
  170. only better.
  171. And as we discover new ways
    to give machines intelligence,
  172. we can distribute that intelligence
  173. to all of the AI assistants in the world,
  174. and therefore to every person,
    regardless of circumstance.
  175. And that is why,
  176. every time a machine gets smarter,
  177. we get smarter.
  178. That is an AI worth spreading.

  179. Thank you.

  180. (Applause)