Transcriptly

Stop Feeding AI for Free | OpenLedger, Ram Kumar

EO09:03Added Aug 31, 2026

Where does AI’s knowledge come from, and who gets rewarded for it? Trillion-dollar models are being built, but the contribution and compensation of everyday data providers are not yet clearly recognized. As we learned

Watch on YouTube →
Contributed by 刘嘉琪

Transcript

Transcript format
  1. 00:00I learned the world through internet. I
  2. 00:02think for my daughter, they will learn
  3. 00:03the world through AI. A company called
  4. 00:05Scale AI, which is known for being a
  5. 00:07data referring company. They've got so
  6. 00:09much data from across the globe.
  7. 00:11People sorting, labeling, and sifting
  8. 00:13reams of data to train and improve AI
  9. 00:16for companies like Meta, Open AI,
  10. 00:19Microsoft, and Google.
  11. 00:20Meta is paying nearly $15 billion for a
  12. 00:23scale AI stake. I've confirmed with a
  13. 00:25source.
  14. 00:25This just shows you how much there is a
  15. 00:27need for data. We all contribute data to
  16. 00:29AI. Your tweet, the posts that you do on
  17. 00:31Facebook, the videos that we upload on
  18. 00:33YouTube are the ones AI models are
  19. 00:35trained on. But us as users, we don't
  20. 00:38get paid for it. The data economy today
  21. 00:40is valued about $1 trillion. This is the
  22. 00:42data. It's across the globe from
  23. 00:44enterprises, organizations, and
  24. 00:46individual people on the internet. And
  25. 00:47let's assume about 5%age of that is
  26. 00:50contributed by individuals. That's close
  27. 00:52to about $500 billion. $500 billion
  28. 00:54worth of data that's taken away from you
  29. 00:56and you're not getting paid for it. We
  30. 00:58want to change that. What if we can flip
  31. 01:00the switch here and have people have the
  32. 01:02ownership to that? We want to build a
  33. 01:04system where you can contribute your
  34. 01:06data and you get paid for that. It's not
  35. 01:08just for AI researchers or developers,
  36. 01:11right? The common man who owns a lot of
  37. 01:13data. We have close to about a million
  38. 01:14users who are contributing data sets for
  39. 01:16that. It's like how nations went ahead
  40. 01:18and fought for oil. Now larger
  41. 01:20organizations are going to fight for
  42. 01:21data, right? Data is the new oil and
  43. 01:24people have to realize that that they
  44. 01:26own that oil. They own that data. But
  45. 01:28it's time to fight back and earn a piece
  46. 01:30of that.
  47. 01:32I'm Ram. I'm one of the core
  48. 01:33contributors at Open Ledger. Open Ledger
  49. 01:35is an AI blockchain where people have
  50. 01:37the data sets. AI systems need these
  51. 01:39data sets. So as an application, you can
  52. 01:41use Open Ledger to go ahead and
  53. 01:43contribute a data set that you own. We
  54. 01:45have lot of data contributors that
  55. 01:47initially came on board. We have close
  56. 01:49to about 10 ecosystem projects building
  57. 01:51AI models on us. We have close to about
  58. 01:53a million users who are contributing
  59. 01:55data sets for that.
  60. 02:03I've been in this industry close to a
  61. 02:04decade right now. The idea was to build
  62. 02:06an R&D company around blockchain and
  63. 02:08machine learning. We saw there is a need
  64. 02:10for enterprises to bring in fairness and
  65. 02:13like transparency within their ecosystem
  66. 02:15like within their organization. We had
  67. 02:16an opportunity to work with enterprises
  68. 02:18like Walmart, Sony, GSK and many more.
  69. 02:20And what we realized is that especially
  70. 02:22a technology like blockchain brings
  71. 02:24equality among every user that uses
  72. 02:26that. Open ledger is a contribution from
  73. 02:29that. Idea was to not just service
  74. 02:32enterprises uh build a product that can
  75. 02:34be used by anyone across the globe and
  76. 02:36figure out how AI is impacting everyone
  77. 02:39lives. We all have this epiphany at one
  78. 02:41point of time where you have
  79. 02:43conversations with your friend about a
  80. 02:44product that you want to buy and you see
  81. 02:46that ad on Instagram you know that your
  82. 02:48data is being used. I've had multiple
  83. 02:50epiphanies as that and we've worked with
  84. 02:52firms where we that is visible right
  85. 02:54people's information was used to make
  86. 02:56their product better sure it gave
  87. 02:57convenience right but it also took
  88. 02:59privacy that's going to happen with AI
  89. 03:01as well AI is going to make money all
  90. 03:03these large organizations going to make
  91. 03:04money out of it but you're not going to
  92. 03:05be part of that as we evolve right as
  93. 03:08models evolve models will become
  94. 03:10specialized where they need data sets
  95. 03:12from people and in that case we need to
  96. 03:14make sure that we can own our data and
  97. 03:16we get paid for it and that's what open
  98. 03:18is trying to solve It's a platform where
  99. 03:20users can come and contribute data sets
  100. 03:22which could be let's say a knowledge
  101. 03:24that they have about trading or a
  102. 03:26knowledge about a particular subject.
  103. 03:28Let's say I know to cook well I can go
  104. 03:30ahead and contribute that and then
  105. 03:32models can use this data and if they use
  106. 03:34that data and they build an AI out of
  107. 03:36that and this AI makes revenue or
  108. 03:39creates an impact you should be a part
  109. 03:41of that you should get a piece of that
  110. 03:42revenue. We want to bring in famas to
  111. 03:44this ecosystem. The people who
  112. 03:46contribute data or a model developer or
  113. 03:48a comput provider or any kind of
  114. 03:50resource provider gets paid as part of
  115. 03:52the process and that's what opener is
  116. 03:54all about. A lot of people say that AI
  117. 03:57is going to make people lose jobs. I
  118. 03:59don't think so. It might be a temporary
  119. 04:01thing but it's going to create a lot of
  120. 04:02jobs. Data contribution itself could be
  121. 04:04a great gig economy. Data is also very
  122. 04:07relevant with enterprises. Data you find
  123. 04:09on internet is very generalized but the
  124. 04:11data that you would find in a firm in an
  125. 04:13enterprise is very specialized right and
  126. 04:15all of his knowledge did not come on the
  127. 04:17internet right people don't write blogs
  128. 04:19about it a surgeon doesn't write about
  129. 04:21how what his experience in actually
  130. 04:23doing the surgery an artist doesn't
  131. 04:24write about how he actually painted a
  132. 04:26picture so it's all comes down to the
  133. 04:28individual person's knowledge that they
  134. 04:30own this knowledge would be needed for
  135. 04:32AI right for AI to truly get into all
  136. 04:35parts of our lives more than just a
  137. 04:36chatbot it needs to no knowledge about
  138. 04:39the entire world. It needs to know about
  139. 04:40very intricate details. Right? In that
  140. 04:42case, people would reach out to you know
  141. 04:45individual users to get their knowledge.
  142. 04:47So we knew that chart GPD moment will
  143. 04:49not going to be there as just a spark.
  144. 04:51It's going to become much more bigger.
  145. 04:52We're going to build AI systems that are
  146. 04:54very specialized in various use cases.
  147. 04:56But there is not enough data out there
  148. 04:58on the internet. We've seen lot of
  149. 05:00independent developers who have lot of
  150. 05:02innovative ideas who are actually
  151. 05:03building interesting AI models in the
  152. 05:05Asian region. They contribute data sets
  153. 05:07by using our nodes which is basically a
  154. 05:10node they can download and have it as a
  155. 05:11plug-in. They can contribute data sets
  156. 05:13for that. If you take a look at uh web
  157. 05:153's nature internet was supposed to be
  158. 05:17decentralized but because of convenience
  159. 05:19we let larger organizations take that.
  160. 05:22So making sure that it does not go back
  161. 05:24to bunch of centralized larger firms is
  162. 05:26very important. We don't have money to
  163. 05:27fight for it. All we have is our own
  164. 05:30power people coming together and
  165. 05:33building something against the larger
  166. 05:34organizations. In order for that to
  167. 05:36happen, you need to make people to come
  168. 05:37together and community building is very
  169. 05:40important as part of that. Building a
  170. 05:42very strong culture, building a very
  171. 05:44strong community. Having the same goal,
  172. 05:46building systems that are open,
  173. 05:48verifiable, and rewarding is what makes
  174. 05:51people come together. Let's take an
  175. 05:52example of Ethereum itself. Ethereum is
  176. 05:54a very communitydriven blockchain and
  177. 05:57that is why it's so strong today. Even
  178. 05:59though it has its ups and down, uh
  179. 06:01Ethereum as an ecosystem is very strong.
  180. 06:03That's why I think community is very
  181. 06:05important in what we're building. So to
  182. 06:08explain uh proof of attribution in a
  183. 06:09very simple manner, what if there's a
  184. 06:11tracking mechanism where you can see and
  185. 06:13who actually contributed for all of this
  186. 06:15and you can also see it on chain that
  187. 06:17everyone who contributed are getting
  188. 06:19paid for it. So it's a tamper-proof
  189. 06:21record of your data's ownership. It's a
  190. 06:24record of how your data was used and
  191. 06:26it's also a record of how much you're
  192. 06:27going to get paid if your data was used
  193. 06:29as well. So the reason why we have the
  194. 06:31proof of attribution is because we need
  195. 06:33to have a trustless system where they
  196. 06:34don't have to believe RAM right they can
  197. 06:37believe the system they can believe the
  198. 06:38code I think that's very important as a
  199. 06:40data contributor you can go ahead and
  200. 06:42choose the model that needs the data and
  201. 06:45you can start contributing data sets to
  202. 06:47that it's that cyclic uh ecosystem you
  203. 06:49want to build all the data that is
  204. 06:51contributed is record on the blockchain
  205. 06:53because so that we can track this and we
  206. 06:54can pay this as a user I know that I can
  207. 06:57prove my ownership by having it on chain
  208. 06:59Another interesting side that we have
  209. 07:01started to see a smaller model
  210. 07:03developers and innovative you know
  211. 07:05people who want to build something
  212. 07:06interesting have started using our
  213. 07:08product. A good example that's being
  214. 07:09built on open edger is a bunch of
  215. 07:11doctors are building uh a sleep model.
  216. 07:13The model is trained on sleep data sets
  217. 07:16high quality sleep data sets across the
  218. 07:18globe. Uh they want to get access to the
  219. 07:20sleep data sets from various parts of
  220. 07:22the world so they can cover various
  221. 07:24races. This particular data set is very
  222. 07:26unique because they're going to
  223. 07:27correlate that with their health vitals
  224. 07:29and then once the model is ready you can
  225. 07:32just upload your sleep data and then
  226. 07:33it's just going to tell you what's your
  227. 07:35body vital looks like. So like this we
  228. 07:37have very interesting players who are
  229. 07:38building models on us very niche
  230. 07:40innovative ideas where you need data
  231. 07:42which is quite unique. We are quite
  232. 07:44excited about them. So a lot of people
  233. 07:46ask me like why this has to use
  234. 07:47blockchain it could be a traditional AI
  235. 07:49company but I don't think so. If you
  236. 07:50take a look at generalized models, the
  237. 07:52era of that will slowly fade away.
  238. 07:54Agents will become much more
  239. 07:56specialized. There'll be an agent for
  240. 07:57healthare. There'll be an agent for
  241. 07:58legal. There'll be agents for every
  242. 08:00other sector out there. And you can have
  243. 08:01a general model power that you need to
  244. 08:03have a specialized model that powers it.
  245. 08:05And the data set is among actually
  246. 08:07people. And these data sets can
  247. 08:09eventually become models, specialized AI
  248. 08:11models which then can be consumed by
  249. 08:14apps and agents that are going to be
  250. 08:15built on top of that. So I think every
  251. 08:17aspect of our life will change. How we
  252. 08:20take a ride home, our doctor visits and
  253. 08:23what we learn from all of that will
  254. 08:25change. I learned the world through
  255. 08:27internet. I think for my daughter AI
  256. 08:29will create a huge impact, right? They
  257. 08:31will learn the world through AI. So that
  258. 08:33AI has to be responsible. So building a
  259. 08:35responsible AI system is upon us. If we
  260. 08:39encourage larger ecosystems to go ahead
  261. 08:41and consume our data and not reward us,
  262. 08:43then that's what is going to happen.
  263. 08:45Right? I think now it's time to change
  264. 08:46it. If we can realize how important our
  265. 08:49data is, probably that's the best output
  266. 08:51that we can see out of AI.