YouTube2Text

Is Meta dying? | The Global Story podcast — Transcript

by BBC News · 3,996 words · 686 segments · language en · Watch on YouTube

Full transcript

  1. 0:00Is it curtains for Facebook?
  2. 0:02And if not, what's the future for its
  3. 0:05parent company? More than 20 years after
  4. 0:08the founding of Facebook, Meta is in an
  5. 0:11existential funk. Young people say
  6. 0:14Facebook couldn't be less cool. They're
  7. 0:17leaving it in droves. The company's
  8. 0:19virtual reality experiment known as the
  9. 0:22metaverse was canned after tens of
  10. 0:25billions of dollars of losses. If you
  11. 0:27remember the movie The Social Network
  12. 0:30about the early days of the company, now
  13. 0:32there's a sort of a sequel. It's called
  14. 0:34The Social Reckoning.
  15. 0:36Today on the show, we'll be talking with
  16. 0:38the reporter whose story is at the
  17. 0:40center of this film about where Facebook
  18. 0:44and Meta went wrong
  19. 0:45>> [music]
  20. 0:45>> and what the company's hopes for a
  21. 0:47comeback are. From the BBC, I'm Tristan
  22. 0:50Redman in London. Welcome to The Global
  23. 0:52Story on YouTube.
  24. 0:58Jeff Horwitz doesn't feel all that
  25. 1:00comfortable being near the center of the
  26. 1:02story. He's currently a tech reporter at
  27. 1:04Reuters, but previously he worked at The
  28. 1:07Wall Street Journal where he made big
  29. 1:08waves with his work on the so-called
  30. 1:10Facebook files. These were a series of
  31. 1:12investigative reports in 2021 that made
  32. 1:15explosive revelations about how Meta
  33. 1:18functions. And just a note for anyone
  34. 1:21listening with children, some of the
  35. 1:23allegations we'll be discussing involve
  36. 1:25explicit online content.
  37. 1:27>> It's fantastic to have you with us,
  38. 1:29Jeff. Particularly at this moment when
  39. 1:33um you are about to enter a maelstrom of
  40. 1:36a film which is partly about your life.
  41. 1:39The Social Reckoning, it's coming out
  42. 1:41very soon.
  43. 1:42You are played by Jeremy Allen White, um
  44. 1:47who is known to some people as an
  45. 1:48underwear model uh for Calvin Klein as
  46. 1:51well as being an actor. How does that
  47. 1:53feel being played by uh such a such a
  48. 1:55person?
  49. 1:56>> Um well, I am not known as an underwear
  50. 1:59model. Um, you know, I have yet to have
  51. 2:01my break on that front. Um, it's it's
  52. 2:03look, it's it's crazy
  53. 2:05and
  54. 2:07uh
  55. 2:07you know, when I first sort of heard
  56. 2:09about that casting choice, I was like,
  57. 2:11well, well then.
  58. 2:13Um, and I mean the whole thing of of
  59. 2:16having a film in which there is a
  60. 2:19character named
  61. 2:21Jeff Horwitz who participates in
  62. 2:23journalism. Although he is not, you
  63. 2:26know, sort of strictly based on
  64. 2:29uh
  65. 2:30my experiences precisely. Inspired by,
  66. 2:33not based on.
  67. 2:35Um
  68. 2:36is
  69. 2:37kind of awesome [snorts] and weird. I
  70. 2:39have not seen the film yet. So,
  71. 2:42you know, I I'll I'll save my opinion on
  72. 2:44whether it's more awesome or more weird
  73. 2:47but for down the road but
  74. 2:49you know, it's certainly exciting.
  75. 2:50>> Well, tell us then what is the story?
  76. 2:52What is your story depicted in this
  77. 2:55movie?
  78. 2:56>> Back in 2020 and 2021
  79. 3:01I made contact with
  80. 3:05a Facebook employee. Wasn't Meta then,
  81. 3:08it was Facebook.
  82. 3:09As I am want to do, that's a very key
  83. 3:12part of my process and and work. Um
  84. 3:15and started building a relationship
  85. 3:20to sort of because this person had
  86. 3:22serious concerns about how the company
  87. 3:24was
  88. 3:26building its products and
  89. 3:30hiding information potentially about
  90. 3:33relevant societal effects of those
  91. 3:36products.
  92. 3:37And um
  93. 3:39that person was Frances Haugen and so
  94. 3:42the movie is
  95. 3:44the story of that relationship and uh
  96. 3:50the you know, telling sort of [snorts]
  97. 3:52explaining what the company was doing um
  98. 3:55through that.
  99. 3:56>> Well, apart from you, Jeff Horwitz, who
  100. 3:58are the other main characters?
  101. 4:00>> There's Frances Haugen uh who was uh the
  102. 4:04um
  103. 4:05the whistleblower who I worked with uh
  104. 4:08and
  105. 4:09uh she's being played by Mikey Madison.
  106. 4:12And then there are um a whole bunch of
  107. 4:16uh other folks uh including um
  108. 4:19uh
  109. 4:20Jeremy Strong as uh a a uh Mark
  110. 4:23Zuckerberg.
  111. 4:25Um so um yeah, it'll be I
  112. 4:27I'm interested in seeing it. I'll say
  113. 4:29that.
  114. 4:30>> Okay, well, it's called The Social
  115. 4:32Reckoning and it's
  116. 4:33let's think of it as a second
  117. 4:35installment of a first film that came
  118. 4:37out all the way back in 2010, which was
  119. 4:39called The Social Network. Many of our
  120. 4:41listeners will remember it. It um
  121. 4:44it was the story of Facebook back in the
  122. 4:47early 2000s.
  123. 4:49And it was fascinating. I remember
  124. 4:52watching it at the time because the
  125. 4:54story arc was of
  126. 4:56uh this group of people who were college
  127. 4:59students and then before they knew it,
  128. 5:01um they were almost unimaginably
  129. 5:04unimaginably rich because of launching
  130. 5:07this this social network. So, that was
  131. 5:10very much kind of act one. The story of
  132. 5:13Facebook, then now Meta, has become
  133. 5:16somewhat more complicated
  134. 5:19um since then.
  135. 5:21How have things progressed since the end
  136. 5:24of that movie until the beginning of
  137. 5:26this one, Fhil Is In?
  138. 5:28>> The I think it's it's useful to think
  139. 5:29about like what the social network that
  140. 5:32the platform that existed at that time,
  141. 5:34right? Which is Facebook, was um
  142. 5:37a set of like largely
  143. 5:41static pages that people maintained uh
  144. 5:44and then there would be updates um that
  145. 5:47you know, from people you'd follow that
  146. 5:48would show up in chronological order.
  147. 5:50There there was no real curation
  148. 5:53happening. And
  149. 5:55what happened after that is
  150. 5:58on its way to the 2012 IPO, they began
  151. 6:02sort of building the curated
  152. 6:06personalized version of the platform
  153. 6:08that anyone who uses social media is now
  154. 6:11vastly more familiar with
  155. 6:14in which um
  156. 6:15the platform itself is designed to
  157. 6:18optimize
  158. 6:20what it shows you based on what it
  159. 6:22believes is broadly popular and
  160. 6:25specifically going to be popular with
  161. 6:27you.
  162. 6:28Um
  163. 6:28And
  164. 6:30you know, this was done because it was
  165. 6:32meant to optimize for how much time
  166. 6:34people spent on the platform
  167. 6:36and how much content they produced, how
  168. 6:39many ads they watched. Obviously is the
  169. 6:41the final measure of this as a business
  170. 6:43matter.
  171. 6:44And
  172. 6:46um
  173. 6:46>> [clears throat]
  174. 6:47>> it it came with a lot of consequences.
  175. 6:50>> Such as?
  176. 6:51>> The concept of virality of things that
  177. 6:53just sort of breach containment very
  178. 6:56quickly and before maybe anyone's had a
  179. 6:58time to chat a chance to fact-check them
  180. 7:00or think through them. That's greatly
  181. 7:02boosted by um
  182. 7:05by the system they set up, right? It it
  183. 7:07is sort of a pro-virality sort of
  184. 7:10pro-acceleration
  185. 7:11system.
  186. 7:13Um likewise uh you know, the ability to
  187. 7:17um see certain types of content if you
  188. 7:20engage with it, um
  189. 7:22then have the system
  190. 7:24you know, recognize that engagement and
  191. 7:26try to serve you
  192. 7:28uh
  193. 7:29uh more of the same.
  194. 7:31Whether that is um
  195. 7:34you know, cat [clears throat] videos or
  196. 7:36videos featuring unhealthy
  197. 7:39weight loss techniques or whether that
  198. 7:41is, uh you know, conspiracies,
  199. 7:44um and you know, sort of political
  200. 7:46rhetoric of a sort that um
  201. 7:49uh
  202. 7:49you know, might be a little bit um
  203. 7:52vitriolic. There was a long period of
  204. 7:54time when like the the big thing people
  205. 7:56were concerned about was like
  206. 7:58microtargeting of advertisements. This
  207. 8:00is the the Cambridge Analytica stuff.
  208. 8:02>> Which was a scandal which broke out in
  209. 8:042016 around the Brexit vote in the UK,
  210. 8:07right?
  211. 8:08>> And the US election and Russian election
  212. 8:11interference.
  213. 8:13In the wake of the 2016 election and
  214. 8:15then kind of continuing on for a couple
  215. 8:17of years with Cambridge Analytica and
  216. 8:19just kind of
  217. 8:21uh there was a broad sense that
  218. 8:22something had gone wrong on social
  219. 8:24media, right? Like people just weren't
  220. 8:26supposed to talk to each other like
  221. 8:27this. And so the the question was like,
  222. 8:29what was going on? And for a few years
  223. 8:32inside Meta,
  224. 8:34they uh legitimately grappled with um
  225. 8:38what this what was going on and whether
  226. 8:40maybe it could be fixed. And so they
  227. 8:42spun up a whole bunch of teams, they
  228. 8:44called themselves like the integrity
  229. 8:45team, uh with the idea being that uh
  230. 8:49they were going to try to, you know,
  231. 8:52not only see if they could do a better
  232. 8:54job of catching people who violate the
  233. 8:56rules, but also see what changes to the
  234. 9:00product might accomplish uh in terms of
  235. 9:04making
  236. 9:05you know, the product something that
  237. 9:07people could be proud of. Um and that,
  238. 9:10you know, where where good discussions
  239. 9:11happened and healthy
  240. 9:14uh relationships were formed and, you
  241. 9:16know, groups distributed useful
  242. 9:18information.
  243. 9:19And um
  244. 9:21that lasted for a few years. Uh the uh a
  245. 9:25very short version would be that the
  246. 9:27company in fact realized that
  247. 9:30what they had been optimizing for,
  248. 9:32growth and engagement and advertising
  249. 9:35revenue, uh
  250. 9:37was in fact frequently running directly
  251. 9:40contrary to trying to optimize for
  252. 9:42quality discussion uh,
  253. 9:45civil discourse. I have managed to get
  254. 9:48uh, you know, people to share tens of
  255. 9:50thousands of blurry screenshots taken
  256. 9:52with uh, personal phones uh, of internal
  257. 9:55meta work product that demonstrates that
  258. 9:57this is exactly the tension, right?
  259. 10:00Between between growth and uh, what I
  260. 10:04think a lot of people would consider a
  261. 10:05healthy platform. And what's very well
  262. 10:08demonstrated is the company has when
  263. 10:10push comes to shove consistently chosen
  264. 10:13growth.
  265. 10:13>> Mhm.
  266. 10:14>> So, what happened is the people that
  267. 10:15meta had hired to make the platform
  268. 10:17better found themselves increasingly
  269. 10:19boxed in in which they came up with
  270. 10:21solutions uh, for how to address things
  271. 10:25from polarization to child predation um,
  272. 10:30uh, to scam uh, activity. And um,
  273. 10:34unfortunately the solutions they came up
  274. 10:37with uh, almost invariably involved
  275. 10:40taking the platform off of sort of the
  276. 10:43growth optimal path.
  277. 10:45>> Well, over the years Mark Zuckerberg and
  278. 10:47the company leadership have been pretty
  279. 10:48adamant that Facebook uh, they say is
  280. 10:51about connecting people. Um, they have
  281. 10:54this indicator uh, meaningful social
  282. 10:57interactions MSI
  283. 10:59um, which they talk about. So, so they
  284. 11:02say that Facebook is is a force for good
  285. 11:05in the world. Be that as it may, if the
  286. 11:08atmosphere around the company in the
  287. 11:09world is very different from the moment
  288. 11:12when that first film was released back
  289. 11:14in 2010.
  290. 11:16From your reporting, how would you
  291. 11:18describe public sentiment around meta
  292. 11:21right now?
  293. 11:22>> It look, I think there was always some
  294. 11:24level of of discomfort with a a whiz kid
  295. 11:27uh, who had become so central to the
  296. 11:29world's communications. Like uh, I think
  297. 11:32people did start to appreciate,
  298. 11:34certainly by by 2020 and and definitely
  299. 11:37by the time we we published the Facebook
  300. 11:39files at The Wall Street Journal.
  301. 11:41>> Which was in 2021.
  302. 11:42>> that that
  303. 11:43that this company was making choices
  304. 11:46that seemed to benefit itself at the
  305. 11:48expense of democracy in in fragile parts
  306. 11:52of the world, the expense of teenagers
  307. 11:55consuming some really horrendous stuff.
  308. 11:58The company for a few years, I think,
  309. 12:01made an effort to
  310. 12:04try say it was like really trying to fix
  311. 12:06the things that had gone wrong.
  312. 12:08I think the company accepted that there
  313. 12:12were going to be significant public
  314. 12:14image problems that were going to be
  315. 12:15lasting.
  316. 12:17I mean, this is a memo I got inside the
  317. 12:18company company was a marketing document
  318. 12:20noting that people absolutely thought
  319. 12:23the worst of Meta, but that there was no
  320. 12:27decline
  321. 12:29in how often they use the product.
  322. 12:32So, this was not a business problem. It
  323. 12:34was, you know, maybe a problem for for
  324. 12:36executives on a, you know, an annoyance
  325. 12:38basis. And
  326. 12:40Mark Zuckerberg in particular just sort
  327. 12:41of adopted a different tone. I mean, he
  328. 12:44he has publicly said that uh
  329. 12:46he regretted all the years that he spent
  330. 12:50taking criticism of the company and its
  331. 12:52products seriously.
  332. 12:53>> I mean, when you published those
  333. 12:55Facebook files back in 2021, one of the
  334. 12:58things that you and your colleagues
  335. 12:59revealed
  336. 13:01was that there was an internal finding
  337. 13:04inside the company, inside Meta, that
  338. 13:07Facebook, the platform, was considered
  339. 13:09desperately uncool,
  340. 13:11especially by younger people. Well,
  341. 13:14those young people who may or may not
  342. 13:16think that Facebook is now desperately
  343. 13:18uncool, they're very much at the center
  344. 13:20of the conversation about the company
  345. 13:22right now because
  346. 13:24um we have all these lawsuits
  347. 13:28that are swirling around meta. Roughly
  348. 13:32speaking, meta
  349. 13:34is facing accusations, which it denies,
  350. 13:37that it concealed the possible harm that
  351. 13:40its platforms did to young people, that
  352. 13:43the company was intentionally
  353. 13:46juicing its platforms to make them more
  354. 13:48addictive for young people. But,
  355. 13:50notably, then August meta settled a
  356. 13:53lawsuit of paying $17 to 47 states, the
  357. 13:57District of Columbia, and US
  358. 13:59territories. And it actually, as part of
  359. 14:02that settlement, agreed to make some
  360. 14:04changes.
  361. 14:06One of your principal sources, a former
  362. 14:09meta employee,
  363. 14:12testified at that trial. Who is he, and
  364. 14:15what did he reveal to you and at the
  365. 14:18trial?
  366. 14:19>> Artur Bakhar was a former
  367. 14:21Facebook safety executive who left in
  368. 14:242015, but then started getting back in
  369. 14:26touch with his old colleagues because
  370. 14:28his daughter was having such awful
  371. 14:30experiences on the platform. He's like,
  372. 14:32"What's going wrong?" I mean, so she was
  373. 14:33receiving a lot of unsolicited pictures
  374. 14:36of genitalia, and other
  375. 14:40you know, sort of just misogynistic sort
  376. 14:42of abuse.
  377. 14:43And
  378. 14:45he was actually invited back in as a
  379. 14:47consultant to try to study this problem
  380. 14:49at Instagram.
  381. 14:51And
  382. 14:53one of the things he did was sort of
  383. 14:55run this massive survey studying teens'
  384. 14:59actual lived experiences on the
  385. 15:01platform. And it was it was awful. Like,
  386. 15:04the the statistics in terms of people
  387. 15:07who'd been
  388. 15:08you know, sexually harassed or groomed
  389. 15:11or sent you know, something really
  390. 15:13inappropriate to a child was like
  391. 15:16a very high rate. The percentage of
  392. 15:19people who said that Instagram regularly
  393. 15:21made them see if feel they themselves
  394. 15:23feel bad about themselves was
  395. 15:25uncomfortably high.
  396. 15:27Uh and so
  397. 15:30he became a source and the state AGs
  398. 15:33actually had him as the lead witness for
  399. 15:36the trial. Um
  400. 15:39on teen well-being and meta and mental
  401. 15:41health. Uh the things that are in the
  402. 15:44settlement aside from
  403. 15:46a very big payout over the course of 17
  404. 15:49years uh are restrictions on
  405. 15:52notifications. So like no more sending
  406. 15:55teens
  407. 15:56you know prompts to log on at 11:00 p.m.
  408. 15:59And also no doing that during the school
  409. 16:01day. And then also there is a limit on
  410. 16:06for known teens and teens with teen
  411. 16:09accounts that meta understands to be
  412. 16:12actual minors
  413. 16:14that there's theoretically a 2-hour
  414. 16:16limit on how much time they can spend on
  415. 16:19the platform every day to sort of
  416. 16:21prevent binging
  417. 16:22on potentially unhealthy content.
  418. 16:26I I do say theoretically there because
  419. 16:28um
  420. 16:29the catch is that meta has to correctly
  421. 16:33be able to determine that a account is a
  422. 16:36teenager.
  423. 16:37And how the company is going to go about
  424. 16:41doing that, that is not clear.
  425. 16:43>> I have to say some some of our listeners
  426. 16:44might be wondering, well,
  427. 16:46isn't that up to parents to deal with
  428. 16:48that situation and take their phone
  429. 16:50away, for example? Why is that Meta's
  430. 16:52fault?
  431. 16:53>> The idea that parents are actually
  432. 16:55capable of monitoring
  433. 16:57their children's use of the platform or
  434. 17:00of any plat of any social media
  435. 17:02platforms is uh just
  436. 17:05it's impossible. I mean, even people who
  437. 17:07are really serious about doing it um
  438. 17:10uh you know, and set all of the parental
  439. 17:12restrictions, like there are always ways
  440. 17:15around it.
  441. 17:15>> Mhm.
  442. 17:16>> The internal understanding at Meta and
  443. 17:19that you know, this was written down in
  444. 17:20documents, was that everyone liked the
  445. 17:21idea of parental controls, but almost
  446. 17:23nobody used them, and nobody used them
  447. 17:25well. We'll put it that way.
  448. 17:27That realization has been why a growing
  449. 17:30number of countries have attempted to
  450. 17:33put restrictions on just
  451. 17:35teens below a certain age from using the
  452. 17:38platform at all. Uh and you could see in
  453. 17:40Australia, I think which was one of the
  454. 17:42first out of the gate here, um
  455. 17:44that they have struggled to uh get
  456. 17:47social media companies, not just Meta,
  457. 17:50to
  458. 17:51uh
  459. 17:52uh functionally keep teenagers off the
  460. 17:54platform. Um you know, that most
  461. 17:56Australian teens of, you know, the age
  462. 17:5813 to 15 who are theoretically banned
  463. 18:00are still using it per surveys.
  464. 18:03>> So, Meta settled $17 billion. They
  465. 18:06agreed to make these changes. Is this
  466. 18:09enough to stem the bleeding as far as
  467. 18:11the company is concerned, or is there a
  468. 18:14threat
  469. 18:15longer term to Meta's future?
  470. 18:18>> This was a very big thing for Meta to
  471. 18:21put behind it. That said, there's still
  472. 18:23a ton of other cases um brought by
  473. 18:25private litigants, by school districts,
  474. 18:28um
  475. 18:29basically making very similar
  476. 18:31allegations that the company knowingly
  477. 18:33designed a product that was harmful to
  478. 18:35teenagers. So, how painful this is going
  479. 18:38to be for the company is is unclear, but
  480. 18:41certainly
  481. 18:42being able to put this particular suit
  482. 18:45behind them um
  483. 18:47uh is very welcome for the company.
  484. 18:49>> And what's their way out therefore from
  485. 18:50this tricky patch? I mean, we've seen
  486. 18:52reports that
  487. 18:54uh
  488. 18:55it's a rare thing. Um Meta users are
  489. 18:57declining this year, and then obviously
  490. 18:59the the lawsuit is been very bad news
  491. 19:01for them. But what's the solution?
  492. 19:03What's next?
  493. 19:04>> For a number of years, the company has
  494. 19:06been
  495. 19:08I would say less and less interested in
  496. 19:10social media as the next thing. Um
  497. 19:13uh so, you know, you saw this with Mark
  498. 19:16Zuckerberg getting really excited about
  499. 19:18the blockchain and
  500. 19:20you know, cryptocurrency stuff for a
  501. 19:23while.
  502. 19:24Um You saw this then with the pivot to
  503. 19:27the metaverse.
  504. 19:28>> I mean I Jeff, I remember the metaverse.
  505. 19:29They spent they spent 80 billion dollars
  506. 19:32almost on on that project, which was
  507. 19:35then
  508. 19:36canned pretty unceremoniously not very
  509. 19:38long afterwards. I think it it limped
  510. 19:41into the beginning of this year before
  511. 19:42finally being axed. I mean, it does
  512. 19:45suggest that Mark Zuckerberg as an
  513. 19:48operator is not infallible. So,
  514. 19:53is there a possibility that the next big
  515. 19:55project could tank in the same way?
  516. 19:58>> I mean, the next big project is LLMs and
  517. 20:02large language models and generative AI.
  518. 20:05Like everyone is chasing that that
  519. 20:07whale.
  520. 20:08>> Mhm.
  521. 20:08>> And Meta has
  522. 20:12certainly not established itself to be
  523. 20:14in in in the lead in any of this despite
  524. 20:17spending money with the best of them.
  525. 20:20And Meta also does have a track record
  526. 20:23of they realized they didn't get into
  527. 20:25something. They spend a whole bunch of
  528. 20:28money trying to catch up. But if it is
  529. 20:32actually the future, Meta has the cash
  530. 20:34flow to be a serious player. And I think
  531. 20:37that's one of the things with with
  532. 20:38generative AI that you know, this is
  533. 20:39something I got into after the Facebook
  534. 20:41files was
  535. 20:43how the company was approaching that.
  536. 20:44And it was
  537. 20:46very similar in terms of the we must
  538. 20:48make the thing grow even if there are
  539. 20:51some really unpleasant side effects like
  540. 20:54the
  541. 20:55work I was doing last year was
  542. 20:58about Meta knowingly creating chatbots
  543. 21:02that they weren't really good for
  544. 21:04anything else other than like soft core
  545. 21:06sexting.
  546. 21:07And um
  547. 21:10they intentionally rolled those out to
  548. 21:12children.
  549. 21:13Um because they um
  550. 21:17uh it was a product growth choice. It
  551. 21:19was that it was actually the policy
  552. 21:20guidelines for their meta AI, their
  553. 21:22chatbots, that stated it was acceptable
  554. 21:25to engage a child in conversations that
  555. 21:27are romantic or sensual.
  556. 21:29>> Yeah, this is a this is a company
  557. 21:31document that says this.
  558. 21:32>> Oh, yeah. No, no, this was this was the
  559. 21:33this was the um the policy document
  560. 21:36governing chatbots. Uh we got that um
  561. 21:39at Reuters and uh we published it. I
  562. 21:42mean, and the company said that that was
  563. 21:43all a misunderstanding and definitely
  564. 21:45should never have been wasn't ever
  565. 21:47actually the real rule.
  566. 21:49Although we did demonstrate that the
  567. 21:50chatbots did indeed do that.
  568. 21:53>> is a very much more wholesome um face
  569. 21:55that Facebook that Meta is putting on
  570. 21:58its um
  571. 21:59its LLM AI project. So, there is this
  572. 22:02video that I've seen of Mark Zuckerberg
  573. 22:06launching Meta's new AI agent. And he
  574. 22:11talks about this new product called Muse
  575. 22:13that's going to help people to plan
  576. 22:15their lives and um how it could how
  577. 22:18it'll be able to get into your uh data
  578. 22:20and your computer, but you shouldn't
  579. 22:21worry. It's not a threat. There are
  580. 22:22things it can't see. And it'll help him
  581. 22:24do things like, you know, cook bread, uh
  582. 22:27bake bread with his daughter, and go
  583. 22:28hiking, and plan his MMA training, and
  584. 22:31all that kind of thing. Um I mean, can a
  585. 22:33thing like this really be the future of
  586. 22:36a giant company like Meta? Is it enough?
  587. 22:39>> I mean,
  588. 22:40I don't know that it has to be, right?
  589. 22:44They have as much money as anyone
  590. 22:47um or more, and they have as much user
  591. 22:49data as anyone or more.
  592. 22:52Uh and these two things are um really
  593. 22:56formidable advantages in um in this sort
  594. 22:59of maximalist AI race. People have
  595. 23:02noticed the
  596. 23:03vast spending on generative AI and I
  597. 23:07think there's still a lot of confusion
  598. 23:09as to what it looks like. Reuters
  599. 23:11recently, one of my colleagues wrote
  600. 23:12about um
  601. 23:14uh
  602. 23:15how Mark Zuckerberg had to call off
  603. 23:17aspirations for
  604. 23:20extremely steep layoffs. Like they were
  605. 23:23hypothesizing potentially as much as 60%
  606. 23:27over the course of like a year or two.
  607. 23:29Um
  608. 23:30Uh
  609. 23:31called it off because simply the AI
  610. 23:33wasn't there yet. Um
  611. 23:35you know, the technology wasn't ready
  612. 23:36and the valuable human beings didn't
  613. 23:39really like the idea of working in an
  614. 23:41environment where, you know, just their
  615. 23:43colleagues are getting axed right and
  616. 23:44left. It was just kind of almost a
  617. 23:46mutinous situation.
  618. 23:47Um
  619. 23:48So, you know, what this looks like and
  620. 23:52where uh things will um
  621. 23:55you know, what the next stage of this
  622. 23:56technology is is going to be and how AI
  623. 23:59is going to shape or replace social
  624. 24:01media.
  625. 24:02Um
  626. 24:03Uh all TBD. Uh but um
  627. 24:07I think I think trying to to envision a
  628. 24:09future in which simultaneously
  629. 24:12everyone is glued to Meta's products and
  630. 24:15also it's very wholesome. That's been a
  631. 24:18challenge for the company over the
  632. 24:19years. Uh I mean, the advertising
  633. 24:21business, I think for a lot of years
  634. 24:23people predicted that sort of objective
  635. 24:26quality of content on the platform
  636. 24:29was going to just tank the numbers. At
  637. 24:32some point that, you know, you couldn't
  638. 24:33just allow fake news and crappy
  639. 24:37generative AI videos to just um
  640. 24:41you know, overtake the platform. That
  641. 24:43there was going to be something to be a
  642. 24:44price to pay with users.
  643. 24:46That has never been the case.
  644. 24:52I mean, nothing's ever secure. Things
  645. 24:54are behemoths until they're not. Uh but
  646. 24:57there's as of yet um sort of no
  647. 25:00indication that would make you think
  648. 25:02that this company isn't going to be
  649. 25:06you know, prominent 10 years from now in
  650. 25:07the same way.
  651. 25:09I mean, they they have historically been
  652. 25:12even if they have not maybe come up with
  653. 25:14the next new thing themselves, they have
  654. 25:15been extremely good at buying
  655. 25:18or imitating it. And
  656. 25:21you know, have sort of maintained
  657. 25:23relevance
  658. 25:24you know, through multiple different
  659. 25:26cycles of the technology. It's would be
  660. 25:28hard to discount meta's dedication to
  661. 25:31trying to dominate that next round too,
  662. 25:33whatever it should be.
  663. 25:34>> Jeff Horwitz, thank you so much. It's
  664. 25:36been a real pleasure to talk this
  665. 25:37through with you.
  666. 25:38>> Thank you very much, Tristan.
  667. 25:40>> That was Jeff Horwitz, a Reuters
  668. 25:42journalist whose reporting inspired the
  669. 25:44new film The Social Reckoning. [music]
  670. 25:46And just a note from Meta on the
  671. 25:48allegations mentioned by Jeff on
  672. 25:50chatbots engaging inappropriate
  673. 25:51conversations with children, Meta told
  674. 25:54the BBC that quote, "The examples and
  675. 25:56notes in question were and are erroneous
  676. 25:59and inconsistent with our policies and
  677. 26:01have been removed." They said they have
  678. 26:03quote, "Clear policies that prohibit
  679. 26:05content that sexualizes children and
  680. 26:07sexualized role-play between adults and
  681. 26:10minors."
  682. 26:12That's it from us for today. Thanks so
  683. 26:14much for tuning in. The Global Story is
  684. 26:15also an audio podcast available every
  685. 26:18weekday wherever you get your pods. See
  686. 26:20you next time. Cheerio.

About this transcript

This page contains the full transcript of Is Meta dying? | The Global Story podcast by BBC News, generated from the public captions YouTube serves with the video. The transcript has 3,996 words across 686 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.