1
00:00:00,000 --> 00:00:00,976


2
00:00:00,976 --> 00:00:09,000
Na lisolo moko, Sam Altman, mokambi ya kompanyi ya AI
OpenAI, alimbolaki mikano na ye mpo na molandi ya ChatGPT.

3
00:00:09,000 --> 00:00:15,679
Ekosala ya kosalela na sanza oyo ezali koya. Altman alobaki boye:

4
00:00:15,679 --> 00:00:22,969
„Nazali kondima ete tozali pene na mokili moko oyo moto akoki
kozala na molandi ya ChatGPT oyo ezali kotala écran ya ordinatɛrɛ

5
00:00:22,969 --> 00:00:29,341
na yo ntango nyonso, ezali kolanda makita nyonso oyo ozali
kosangana na yango, ezali koenregistre babɛlɛ nyonso

6
00:00:29,341 --> 00:00:37,279
mpe na ndenge wana ezali na contexte ya kokoka ya
bomoi na yo mobimba – ya makambo nyonso oyo ozali komona.“

7
00:00:37,279 --> 00:00:42,840
Na mwa ntango moke, AI ekosala kaka te
eyano na mituna, kasi ekosala makambo moko mpe.

8
00:00:42,840 --> 00:00:47,703
Ekolakisa makanisi, ekoyeba mikakatano oyo
ekoki kozala mpe ekosunga na bomoi ya

9
00:00:47,703 --> 00:00:51,161
mokolo na mokolo. Mokili ya sika kitoko boni!

10
00:00:51,161 --> 00:00:56,510
Na pembeni na yango, ba dystopie ya kala lokola 1984 – „Big
Brother is watching you“ – ezali komonana

11
00:00:56,510 --> 00:01:02,178
lokola théâtre ya bapopɛ ya kitoko, mpamba te
batali ya Orwell bazalaki kaka sima ya efelo.

12
00:01:02,178 --> 00:01:08,526
Batali ya Altman bazali na smartphone to PC na bango
moko mpe kala mingi te bakozala na bongo na bango moko.

13
00:01:08,526 --> 00:01:12,650
Bakokoma baninga ya ntango nyonso, baninga ya lokuta.

14
00:01:12,650 --> 00:01:15,343
Mopɔti ya bozangi mpasi ezali na mbilinga.

15
00:01:15,343 --> 00:01:23,476
Nani akolinga te kotika ba rendez-vous moko oyo ezali
kosepelisa te to kosunga yango moko na moko ya kokoka na mosala?

16
00:01:23,476 --> 00:01:30,363
Kasi tango bato bakokoma kaka kotɛlɛma na lisungi ya
ba AI ya sika, AI ekobanda kokɔtisa bango na ndingisa.

17
00:01:30,363 --> 00:01:37,520
Nzela moke kati ya „lisungi“ mpe „bokonzi“ ekokoma
te kosalema na basaleli, kasi na ba algorithmes.

18
00:01:37,520 --> 00:01:41,204
Na ndenge wana lisungi ya kosala
makambo ekokoma bokɔti ya bokonzi.

19
00:01:41,204 --> 00:01:47,020
Awa ndakisa moke ya ndenge yango ezali:

20
00:01:47,020 --> 00:01:50,695
1. Moto alingi kokoma maloba ya
komekameka likambo ya politiki ya guvɛrnema?

21
00:01:50,695 --> 00:01:55,437
AI, oyo ezali kolanda bouton nyonso oyo
ebɛtami, „ebongisi“ makomi yango liboso.

22
00:01:55,437 --> 00:02:04,040
Alobi maloba ya ndenge ya kitoko to abimisi yango te,
mpo ekomami lokola «ekosemba te» to «ekaboli bato».

23
00:02:04,040 --> 00:02:11,060
Sikoyo AI ya Google ezali kopimela ba commentaires
ya kitoko mpo na makambo oyo yango moko elimboli.

24
00:02:11,060 --> 00:02:15,078
2. Motu azali koluka ba informations na makambo
oyo bato bazali koswana na yango na internet?

25
00:02:15,078 --> 00:02:19,905
AI ezali kopona ba résultats ya boluki mpe ezali
kolakisa kaka ba sources oyo ezali komona lokola

26
00:02:19,905 --> 00:02:22,414
«ya kokoka» – elingi koloba oyo ekokani na système.

27
00:02:22,414 --> 00:02:25,585
Sikoyo Shadow Banning ezali na ba réseaux sociaux.

28
00:02:25,585 --> 00:02:34,269
Ba contenus ezali ya kokanga te, kasi ezali kolakisama lisusu
na boluki te mpe bongo ezali kozwama lisusu na ba abonnés te.

29
00:02:35,769 --> 00:02:39,908
3. Kotanga mpe kokoma esengeli lisusu te, mpo
AI ezali kosangisa ba textes ya mpasi to

30
00:02:39,908 --> 00:02:42,683
ezali kokoma ba articles mpe ba discours mobimba.

31
00:02:42,683 --> 00:02:49,870
Mokomi Tom-Oliver Regenauer akomaki na
yango kala mingi te na document ya position:

32
00:02:49,870 --> 00:02:56,020
«Motu oyo azali kobimisa kotanga, kokanisa mpe kokoma
na libanda, abosanaka yango – mpe na lombango mingi.

33
00:02:56,020 --> 00:03:01,599
Bongo ya moto ezali lokola musele. Soki e training te, ekokufa...

34
00:03:01,599 --> 00:03:10,504
Soki motu akanisi ndenge oyo bokoti ya smartphone banda 2007
ebimisaki ba conséquences ya mabe, esengeli te imagination

35
00:03:10,504 --> 00:03:19,616
mingi mpo na extrapoler , ndenge oyo AI na tango mokuse
koleka ekosala na ba capacités cognitives ya espèce na biso.

36
00:03:19,616 --> 00:03:26,343
AI – lokola mpe smartphone – ezali liboso
technologie ya militari mpe instrument ya dégénérescence.

37
00:03:26,343 --> 00:03:32,294
Esaleli oyo bato mingi mpenza bakoki kosalela
yango na responsabilité mpe na productivité te.

38
00:03:32,294 --> 00:03:38,814
Ba „Smombies“ ya kokamba te, ya mposa te mpe ya
makanisi te, oyo bazali kaka kobongisa engumba,

39
00:03:38,814 --> 00:03:43,840
ezali lokola ndakisa ya kobangisa kasi eleki te.“

40
00:03:43,840 --> 00:03:49,169
4. AI ezali koyeba na tango mokuse mpenza makanisi
ya moto na politiki, ezali kotanga bansango ya

41
00:03:49,169 --> 00:03:52,821
sekele mpe eyebi boyangeli ya bondeko mpe ya bakontak.

42
00:03:52,821 --> 00:03:58,580
Azali kosalela boyebi oyo te mpo na kobakisa
moto na bokoli na ye, kasi mpo na „kobongisa“ ye.

43
00:03:58,580 --> 00:04:02,526
Elimbolami malamu: Kobongisama na mobeko oyo esengami.

44
00:04:02,526 --> 00:04:06,637
Nudging, elingi koloba bokonzi ya makanisi
na nzela ya mayele, esalelamaki na ntango

45
00:04:06,637 --> 00:04:10,245
ya mbula ya Corona na nzela ya bameya mingi.

46
00:04:10,245 --> 00:04:15,040
Boni ya bokonzi bapesi ChatGPT mpe baninga na ye, biso basaleli?

47
00:04:15,040 --> 00:04:19,198
Tom-Oliver Regenauer apesi maloba ya polele:

48
00:04:19,198 --> 00:04:27,392
„Nazali kolakisa te ete natangisa systeme oyo ezali kotala ngai
na sekele tango nyonso mpe oyo ezali na makoki ya koboma ngai.

49
00:04:27,392 --> 00:04:36,955
Mpo ete Palantir, ChatGPT ya OpenAI, Claude ya Anthropic mpe Grok
ya SpaceXAI bazali kosala yango lelo na Gaza, Ukraine, Iran mpe

50
00:04:36,955 --> 00:04:44,961
na bisika mosusu ya bitumba, epai ba systemes ya bibuneli ya
maboko ya moto ezali kobongisama na ba systemes

51
00:04:44,961 --> 00:04:52,424
yango mpenza – lokola ba mini-drones oyo, na
maloba ya EUROPOL, ekotambola na babalabala na biso.

52
00:04:52,424 --> 00:05:00,370
Moto oyo asaleli ba systemes ya AI, atangisi „SkyNet“,
databank ya mokuse oyo ekobongisa „Terminator“ ya solo.

53
00:05:00,370 --> 00:05:06,784
Mpe yango, na nzela ya Boston Dynamics, Unitree mpe
Tesla, ekokoma liboso ya ndako na biso kala mingi te.“