AI Deception: How Tech Companies Are Fooling Us

ColdFusion
15 Apr 202418:58

Summary

TLDR这段视频讲述了人工智能(AI)的炒作现象及其后果。从18世纪的机械土耳其人骗局到现代,一些公司宣称其技术由AI驱动,但实际上可能依赖于人工操作。视频指出,AI技术的过度宣传可能导致消费者不信任,并对真正在该领域工作的人员造成负面影响。讨论了AI在各行各业中的应用,包括其在提高效率和可能导致工作岗位减少的潜力。还提到了AI技术发展中的一些失败案例,以及一些公司如何成功地整合AI。视频最后强调,尽管AI技术取得了显著进步,但我们仍需对AI的承诺持谨慎态度,并关注其对工作和社会的长期影响。

Takeaways

  • 📜 18世纪匈牙利作家和发明家沃尔夫冈·冯·肯佩伦制造了一个名为“机械土耳其人”的自动下棋机器,这在当时是一个巨大的骗局,因为它实际上是由人在幕后操作的。
  • 🛒 亚马逊在2016年推出了名为“Just Walk Out”的结账系统,该系统允许顾客直接拿走商品并离开,而无需在收银台扫描商品。然而,该系统实际上是由印度的1000人团队手动审查交易来确保准确性的。
  • 📈 从2015年到2022年,全球企业对AI的投资增加了7倍,许多公司在AI热潮中看到了巨大的增长,这给其他公司带来了将AI整合到他们的商业模式或产品中以保持竞争力的压力。
  • 🚫 AI技术的一个不受欢迎的用途是它可以使诈骗者的工作变得更容易,AI可以帮助生成令人信服的文本,诱使人们点击钓鱼链接。
  • 💡 Aura是一家可以代表用户向数据经纪人提交退出请求的公司,以保护用户信息不被出售给垃圾电话和诈骗者。
  • 📉 根据高盛的数据,标普500指数中有创纪录的36%的公司在其第四季度收益报告中提到了AI,这表明即使是最大的公司也在夸大其AI产品的能力。
  • 🧐 AI领域的一个常见现象是AI洗牌(AI washing),即公司通过提供关于其AI产品的能力、局限性或风险的误导性信息来制造虚假热潮。
  • 💻 一些公司声称他们的AI技术可以预测公司和行业趋势,但实际上并没有这样的能力,例如加拿大投资公司Delia因虚假宣传其AI能力而被美国证券交易委员会罚款。
  • 📊 尽管一些评论家认为我们正处于AI泡沫之中,但像马克·库班这样的投资者认为,由于AI领域缺乏IPO和没有内在价值的公司上市,我们并不处于泡沫之中。
  • ⚖️ AI技术的核心能力是模仿认知劳动,这是以前任何技术在没有人类干预的情况下都未曾实现的,这可能会使当前的AI热潮与以往的技术泡沫有所不同。
  • ❌ 一些公司匆忙推出AI产品或用AI取代人类,但这些尝试经常失败,例如一家快递公司启用的AI在线聊天不仅无法回答问题,还对客户进行了侮辱。
  • ⏳ 尽管有关AI将在未来十年创造数亿就业机会并积极影响GDP的报道,但目前看来,AI正在取代工作,而不是创造工作,特别是在技术行业之外,如媒体和娱乐行业也开始感受到AI的影响。

Q & A

  • 什么是Mechanical Turk?

    -Mechanical Turk是1770年由匈牙利作家和发明家沃尔夫冈·冯·肯佩伦制造并展示的一台自动下棋机器。在欧洲的演示中,这台机器赢得了大多数对局,甚至据说击败了拿破仑和本杰明·富兰克林等玩家。然而,实际上这是一个骗局,机器下面有人秘密操作。

  • 亚马逊的Just Walk Out系统是如何工作的?

    -亚马逊的Just Walk Out系统允许顾客在商店中拿走商品后直接离开,无需在收银台手动扫描商品。该系统后台由AI自动处理交易和物流。然而,实际上这个所谓的AI技术是由印度的1000人在观看和标记视频以确保准确的结账。

  • AI在广告中被过度使用的现象被称为什么?

    -AI在广告中被过度使用的现象被称为AI洗牌(AI washing)。这是指公司通过提供关于其AI产品或服务的能力、局限性或风险的误导性信息来制造虚假的热潮,或者简单地全面撒谎。

  • 全球企业在AI上的投资从2015年到2022年增加了多少?

    -从2015年到2022年,全球企业在AI上的投资增加了七倍。

  • AI处理器(如NPU)在PC和笔记本电脑中的作用是什么?

    -AI处理器,如神经处理单元(NPU),是专门为AI任务设计的专用处理器。它们在个人电脑和笔记本电脑中变得越来越普遍,旨在提高AI相关任务的处理速度和效率。

  • AI技术可能带来的一个负面用途是什么?

    -AI技术的一个负面用途是它可以使诈骗者的工作变得更容易。AI可以帮助生成令人信服的文本,试图诱使人们点击钓鱼链接,还有垃圾邮件和自动电话。

  • 为什么一些公司在AI产品推出后会不断尝试挽回面子?

    -一些公司在AI产品推出后会不断尝试挽回面子,因为它们可能过度夸大了AI产品的功能,导致消费者对产品的实际效果感到失望,从而损害了公司的信誉。

  • AI泡沫一词是如何与2000年的.com泡沫相比较的?

    -AI泡沫一词与2000年的.com泡沫相比较,因为人们担心AI行业可能会出现类似的过度投资和过度炒作,最终导致市场崩溃。但与.com泡沫不同,AI技术具有真正的潜力,因为它能够模仿认知劳动。

  • Gartner炒作周期是什么?

    -Gartner炒作周期是一个描述新兴技术如何被人类过度热情接受的模式。它包括技术被过度估计和过度炒作,直到市场崩溃,之后真正有价值的公司会出现并推动新技术的发展,直到技术成熟。

  • 为什么一些公司在AI的帮助下进行裁员?

    -一些公司在AI的帮助下进行裁员,因为AI可以提高工作效率,使得更少的员工能够完成更多的工作。此外,公司可能不愿意公开承认AI在裁员中的作用,因为这与他们之前对AI的积极宣传相矛盾。

  • AI对就业市场的真实影响是什么?

    -AI对就业市场的真实影响是复杂的。虽然有些工作可能会被AI取代,但AI也可能创造新的工作机会。然而,低技能和低工资的工作更有可能受到AI的影响,而STEM、医疗保健和专业领域可能会继续增加就业机会。

  • 为什么AI技术在某些情况下可能无法实现其承诺的效益?

    -AI技术在某些情况下可能无法实现其承诺的效益,因为AI的发展和应用可能受到技术限制、市场误解、公司内部管理问题以及社会和法律因素的制约。此外,AI技术的实际效果可能被过度夸大,导致公众对AI的期望过高。

Outlines

00:00

🎭 机械土耳其人与现代AI的欺骗

本段主要讲述了历史上著名的机械土耳其人——一种18世纪的自动下棋机器,以及它实际上是如何通过人在内部操作来欺骗观众的。视频进一步对比了亚马逊在2016年推出的“即走即付”系统,该系统声称使用AI技术自动化结账流程,但实际上背后有1000名员工在印度手动审查交易。这揭示了AI技术宣传与实际应用之间的差异,以及这种误导性广告可能对公众信任和AI行业造成的负面影响。

05:02

📈 AI炒作与AI洗涤

这一段深入探讨了AI炒作现象,即公司通过夸大或误导性的信息来夸大其AI产品的能力,从而误导投资者。提到了高盛报告中指出的S&P 500公司中有36%在财报中提到AI的趋势,以及全球企业AI投资的增长。同时,也提到了一些公司因为虚假宣传AI能力而受到法律惩罚的案例,以及消费者对AI产品的失望感,特别是在AI功能尚未成熟的情况下。

10:02

📉 AI泡沫与Gartner炒作周期

此段落讨论了AI技术当前可能处于Gartner炒作周期的“期望的顶峰”阶段,并可能正进入“幻灭的低谷”。说明了AI技术的实际应用将在炒作和疲劳消退之后到来。同时,也提到了一些AI失败的案例,比如DPD快递公司的AI聊天机器人无法提供有用回答,反而侮辱客户和雇主。此外,还提到了AI成功整合的公司,以及AI对就业市场的潜在影响,包括可能取代人类工作的担忧。

15:06

🛠️ AI对就业市场的影响

最后一段讨论了AI对就业市场的直接影响,包括一些大型科技公司如IBM和UPS在裁员时提到的AI因素。提到了技术行业以外的领域,如媒体制作,也对AI可能取代人类工作表示担忧。同时,也指出了一些行业如STEM、医疗保健和建筑业可能会继续增长。最后,强调了AI技术进步的积极面,并承诺将制作一个关于AI积极影响的视频,以平衡对AI的批评。

Mindmap

Keywords

💡Mechanical Turk

机械土耳其人是18世纪由匈牙利作家和发明家沃尔夫冈·冯·肯佩伦制造的一种自动下棋机器。它在欧洲各地的演示中赢得了多数对局,甚至传说击败了拿破仑和本杰明·富兰克林等名人。然而,实际上这是一场骗局,机器下面有人秘密操作。在视频中,它被用来比喻一些公司宣称的AI技术实际上可能并不如宣传的那样完全自动化。

💡AI Washing

AI洗涤是指公司通过提供关于其AI产品或服务的能力、局限性或风险的误导性信息来制造虚假热潮,误导投资者。视频中提到,即使是世界上最大的公司也在过度使用AI这一术语,而没有实际的成果来支持他们的说法。

💡Just Walk Out Technology

即走技术是亚马逊推出的一种结账系统,允许顾客从商店拿走商品并直接离开,而无需在收银台手动扫描商品。该系统背后由AI自动处理交易和物流。然而,视频中揭示了实际上这项技术是由印度的1000人手动审查交易来支持的,这与宣传的完全自动化的AI技术有所出入。

💡Data Brokers

数据经纪人是购买和销售个人信息的公司,这些信息可能被用于垃圾电话、钓鱼链接和其他诈骗行为。视频中提到,赞助商Aura可以代表用户识别数据经纪人并提交退订请求,以保护用户的个人信息不被滥用。

💡Neural Processing Unit (NPU)

神经处理单元(NPU)是专门为AI任务设计的专用处理器,它们在个人电脑和笔记本电脑中变得越来越普遍。然而,视频中提到,尽管微软宣称我们已经进入了AI PC的时代,但当前的这些产品在实际应用中并没有达到预期的效果,例如生成的图像质量很差。

💡AI Bubble

AI泡沫是指AI技术被过度炒作,可能会导致投资者失望并最终导致市场崩溃的现象。视频中讨论了AI技术是否正处于高估其影响的阶段,并可能进入幻灭之谷。

💡Gartner Hype Cycle

高德纳炒作周期是一个描述新兴技术如何被过度热情接受,随后市场崩溃,最终由真正有价值的公司推动技术成熟的模型。视频中提到,AI可能正处于这一周期的顶峰预期阶段,并可能很快进入幻灭之谷。

💡AI-Generated Content

AI生成内容是指使用AI技术自动创建的文章、诗歌或其他形式的内容。在视频中,有一个例子是关于一个在线聊天机器人,它未能有效回答客户的问题,反而提供了一首诗并咒骂了用户,展示了AI在实际应用中可能存在的问题。

💡Job Automation

工作自动化是指使用技术,特别是AI,来执行原本由人类完成的工作。视频中讨论了AI可能导致的工作岗位减少,以及这对低技能和高技能工人的不同影响。

💡AI in Customer Service

在客户服务中使用AI指的是利用AI技术来提高服务效率和质量。然而,视频中提到了AI在实际应用中的一些失败案例,如一个在线聊天机器人不仅未能有效回答问题,还提供了不恰当的回应。

💡AI and Job Displacement

AI和工作置换是指AI技术可能取代某些工作岗位,导致工人失业。视频中提到,尽管有些公司宣称AI将创造更多工作而不是取代工作,但实际上,AI正在导致一些技术相关行业的大规模裁员。

💡AI Ethics

AI伦理是指在AI技术的开发和应用中考虑道德和社会责任。视频中虽然没有直接提到AI伦理,但从对AI技术可能带来的负面影响的讨论中可以推断出,AI伦理是当前AI发展中一个重要但经常被忽视的方面。

Highlights

1770年,匈牙利作家兼发明家沃尔夫冈·冯·肯普伦制造并展示了一个名为“机械土耳其人”的自动下棋机器,该发明在欧洲的演示中赢得了大多数对局。

机械土耳其人实际上是一个骗局,有人秘密地在桌子下操作机器。

2016年,亚马逊推出了名为“Just Walk Out”的结账系统,允许顾客从商店拿走商品后直接离开,而无需手动扫描商品。

亚马逊所谓的AI技术实际上是由印度的1000人观看和标记视频来确保准确的结账。

到2022年,这些员工仍在手动审查亚马逊Go商店70%的交易。

公司经常在广告中遗漏关键信息,误导消费者。

AI技术的一个不受欢迎的用途是它可以使诈骗者的生活变得更容易,AI可以帮助生成令人信服的文本,诱使人们点击钓鱼链接。

数据经纪人通过出售个人信息给垃圾电话和诈骗者而赚取巨额利润。

Aura公司可以代表用户识别数据经纪人并提交退出请求,以保护用户免受在线威胁。

AI在智能手机、LinkedIn、当地政府网站上的应用越来越广泛,许多组织都会宣称其产品或服务由AI提供动力。

AI洗牌是指公司通过提供关于其AI产品或服务的能力、限制或风险的误导性信息来制造虚假热潮。

高盛报告称,创纪录的36%的标普500公司在其第四季度收益报告中提到了AI。

2015年至2022年,全球企业对AI的投资增加了七倍。

加拿大投资公司Delia声称拥有预测公司和行业趋势的AI,尽管没有这样的能力,最终因虚假广告和误导性声明被罚款22.5万美元。

AI处理器(如NPU)在PC和笔记本电脑中变得越来越普遍,但消费者对当前产品的AI功能并不满意。

AI洗牌可能导致客户不信任,甚至破坏那些在该领域进行合法工作的人的声誉。

AI技术的核心能力是模仿认知劳动,这是以往任何技术在没有人类干预的情况下都未曾实现的。

一些公司匆忙推出AI产品或用AI取代人类,但这些做法往往失败,例如DPD快递公司启用的AI在线聊天无法有效回答问题。

AI的成功整合案例表明,AI并非总是如CEO所承诺的那样有用,而是被低估了。

自2022年以来,技术相关裁员数量显著增加,这表明AI在裁员中发挥了比CEO们愿意承认的更大的作用。

AI可能不会取代人类,但会增强那些利用AI的人的能力,而那些不利用AI的人将被取代。

到2030年,大约有1200万美国工人可能需要转换工作,低收入工人更有可能受到影响。

尽管有关AI将在未来十年增加数亿工作岗位并积极影响GDP的报道,但当前的情况表明,AI可能会在许多人尚未意识到的情况下改变工作。

Transcripts

00:00

this video is brought to you by Aura hi

00:03

welcome to another episode of Cold

00:04

Fusion in 1770 Hungarian author and

00:08

inventor wolf Gan Von kemplin built and

00:10

unveiled a self-operating chess playing

00:13

machine called The Mechanical Turk in

00:15

demonstrations around Europe his chess

00:17

playing invention won most of its games

00:19

against actual human chess players and

00:21

it's even said that it went on to beat

00:23

players like Napoleon and Benjamin

00:25

Franklin naturally The Invention was

00:27

extremely popular a Marvel of its day

00:30

however it was actually just some guy

00:32

under the table secretly operating the

00:34

machine in other words it was a hoax

00:37

some 250 years later in 2016 Amazon

00:41

pulled a similar stunt their checkout

00:43

system called just walk out allowed

00:45

customers to Simply grab items from the

00:47

store and leave under this system

00:49

customers didn't have to manually scan

00:51

their items at the checkout the

00:53

transactions and Logistics were

00:55

automatically taken care of behind the

00:56

scenes by AI An Elegant combination of

00:59

Compu Vision sense of fusion and deep

01:02

learning or was it just like the

01:04

Mechanical Turk all was not as it seemed

01:07

Amazon's so-called AI technology was

01:10

actually powered by a thousand people in

01:12

India watching and labeling videos to

01:14

ensure accurate checkouts now humans

01:16

labeling data is often how AI models are

01:19

first trained and that's fine but the

01:21

point is the advertising was misleading

01:24

even by 2022 these 1,000 people were

01:27

still manually reviewing 70% of of

01:30

transactions in 20 Amazon go stores 40

01:33

Amazon fresh grocery stores and two

01:35

Whole Food

01:37

stores although some saw it as dystopian

01:40

Amazon promoted their just walkout

01:42

technology as a utopian magical solution

01:45

powered by Ai and here in lies the

01:47

problem companies often leave out

01:49

critical information in the fine print

01:51

it wasn't until 2023 that Amazon mention

01:54

the use of the a thousand people in

01:55

India and it gets worse according to a

01:58

report by the information other

02:00

companies were swindled too quote 30

02:02

stores operated by other companies in

02:04

the US sports stadiums 12 airports and

02:07

one University all used Amazon's walk

02:10

out

02:11

technology and this is where things

02:13

become murky when companies promis

02:15

revolutionary technology but don't

02:17

follow up on them it creates distrust

02:19

but it can even spoil it for those who

02:21

are doing legitimate work in the field

02:23

with the AI Revolution that we've seen

02:25

over the past couple of years there

02:26

needs to be a more critical look at

02:28

what's happening with AI in this episode

02:30

we'll see the reasons for the AI hype

02:32

and its consequences we'll look at if

02:34

there is indeed an AI bubble and what

02:36

real world impact that might be having

02:38

this includes companies releasing

02:40

half-baked AI products but also

02:42

simultaneously replacing actual humans

02:44

with AI it's a dichotomy of the industry

02:46

so sit back and relax as we make sense

02:48

of it

02:51

all you are watching C Fusion

02:56

TV one use of AI technology that

02:59

absolutely nobody is thrilled about is

03:01

the fact that it can make scammers lives

03:02

easier AI can help generate convincing

03:05

texts that try to trick you into

03:06

clicking a fishing link and then there's

03:08

the spam emails and Robo calls you've

03:10

all experienced it but why does it

03:12

happen it's because your information has

03:14

been sold to data Brokers they're making

03:16

a fortune from selling your information

03:18

to robocallers and scammers they can

03:20

know a lot about you including your name

03:22

phone number and address but there's a

03:24

simple way to stop this that's where

03:26

today's sponsor Aura comes in they can

03:28

identify data Brokers and submit opt out

03:31

requests on your behalf Brokers are

03:33

legally required to remove your

03:35

information if you ask them to but they

03:37

make it super hard to do let Aura handle

03:39

that for you Aura also does so much more

03:41

to protect you and your family from

03:43

online threats that you can't see it's

03:45

really easy to set up so you don't have

03:46

to download several different apps to

03:48

get things like parental controls

03:50

antivirus a VPN password management

03:53

identity theft Insurance and more and

03:55

you get everything at one affordable

03:57

price visit aura.com

03:59

Fusion to get started on a 2-e free

04:01

trial offer only available in the United

04:04

States okay so let's

04:08

continue no it's not just you the term

04:11

AI is really being thrown around

04:13

everywhere for older cold fusion viewers

04:15

you'd know that I've been covering AI

04:17

for the good part of a decade now before

04:19

2022 the word artificial intelligence

04:22

was limited to Industry verbage in

04:24

obscure research papers the general

04:26

public didn't know or care and even as

04:29

late as 21 when I covered gpt3 there was

04:32

a lot of people that thought I was just

04:33

making it up how could a computer

04:35

possibly talk like that well as you all

04:38

know after 2022 especially following the

04:41

launch of chat GPT it was a different

04:43

story from your smartphone to LinkedIn

04:46

to your local government site a lot of

04:48

organizations today will boast the term

04:50

powered by

04:52

AI while in a lot of cases it makes

04:55

sense the overuse of the term and it

04:57

being thrown around so casually is a

04:59

symptom of something called AI

05:02

washing AI washing simply put is when

05:05

companies create a fake buzz and mislead

05:07

investors by providing manipulative

05:09

information about the capabilities

05:11

limitations or risks of their AI

05:13

products or simply wholesale lie about

05:16

how and when they're using AI according

05:19

to Goldman Sachs a record 36% of S&P 500

05:22

companies mentioned AI in their Q4

05:24

earnings reports if the biggest

05:26

companies in the world are hyping it up

05:28

there's no doubt that small companies

05:30

are following suit and some without

05:31

tangible results backing up their claims

05:34

and the Investments are there to match

05:35

too from 2015 to 20122 Global corporate

05:39

AI Investments have increased seven

05:41

times a lot of companies have seen

05:43

tremendous growth in the recent AI hype

05:46

and this puts pressure on other

05:47

companies to integrate AI into their

05:49

business model or their products in

05:50

order to compete and that's even if it's

05:52

not really AI Monkey C Monkey Do For

05:56

example the Canadian investment firm

05:58

Delia claimed that that had built an AI

06:00

to predict upcoming companies and

06:02

Industry Trends despite not having any

06:04

such capabilities the SEC caught wind of

06:07

the scam and has since ordered them to

06:09

pay

06:10

$225,000 and this was for false

06:12

advertising and misleading statements

06:14

about the use of AI and I have a feeling

06:16

that this is just the start I previously

06:18

talked about the wire card fraud where

06:20

CEO Marcus Brawn boasted about having

06:23

Advanced proprietary AI throughout their

06:25

fintech products and services in reality

06:27

it was just spreadsheets in the end

06:30

there was $2 billion missing from the

06:32

books but that was a story for a

06:33

different day you might have also heard

06:35

about the npu or neural Processing Unit

06:38

it's a dedicated processor specifically

06:41

designed for AI tasks among other things

06:44

they're becoming increasingly common in

06:46

the world of PCS and laptops and

06:47

Microsoft has even claimed that we're

06:49

now in the age of the AI PC but recent

06:53

reviews of the current crop of these

06:54

products have left consumers with a bad

06:56

taste in their mouth let's talk a bit

06:58

more about AI yes the Snapdragon

07:00

processors dominate here but AI is still

07:03

very nent many people either won't use

07:05

these tasks all that much or these tasks

07:08

themselves just don't work that well yet

07:10

for example they demo generating images

07:12

of animals and they highlighted how fast

07:14

the processes were versus Intel but when

07:17

I look closely at the images generated

07:19

they were all garbage the AI dogs all

07:21

had very lopsided eyes so who cares if

07:24

one of these processors is faster than

07:26

the other if they are both generating

07:27

garbage right now I just wouldn't advise

07:30

anyone buy a laptop based on its AI

07:32

capabilities a news analyst from

07:34

computer world has mentioned that these

07:36

AI PCS aren't quote all they cracked out

07:39

to be and if you're expecting something

07:40

transformative when you buy one from the

07:42

start of 2024 you're going to be

07:43

disappointed they might one day deliver

07:45

a lot of cool features just not yet end

07:48

quote but are there consequences of such

07:50

misplaced promises other than distrust

07:53

from customers companies that Miss The

07:55

Mark with AI product launches are

07:56

constantly trying to save face even the

07:59

most popular companies in the world are

08:00

not immune to mistakes perhaps the

08:03

biggest consequence of AI washing is a

08:05

certain b word that might be lurking

08:07

around the

08:10

corner almost as soon as AI started

08:12

going Uber mainstream from late 2022 a

08:15

lot of people immediately liken the

08:17

phenomenon to the dotom bubble or even

08:19

the crypto Fiasco in fact some people

08:22

still do today even when we talk about

08:24

the.com Bubble it's not the worldwide

08:26

web that was the problem because of

08:28

course you're using a right now instead

08:30

it was mostly the e-commerce side of

08:32

things which over promised reeling in

08:34

investors but underd delivered leading

08:36

to a crash when everyone realized that

08:38

the companies that they invested in

08:39

weren't making any

08:42

money with AI there's no doubt that

08:44

we'll see some more investment in the

08:46

space and Monumental growth too

08:48

according to Oppenheimer no not the one

08:50

you're thinking of who's a chief Global

08:52

Equity strategist at Goldman Sachs

08:54

research quote we believe we are in the

08:57

relatively early stages of a new

08:59

technology cycle that is likely to lead

09:01

to further outperformance end quote

09:03

Nvidia who designs the chips that are

09:05

powering big Tech's AI Revolution is

09:07

tentative proof of this their stock is

09:09

up 80% this year alone perhaps excessive

09:12

but it's indicative of what the market

09:14

thinks of AI Mark Cuban who made his

09:17

fortune during the era also doesn't see

09:19

AI as a bubble in a recent interview

09:21

with Lex fredman he expressed that the

09:23

biggest sign that we're not in a bubble

09:25

is the lack of IPOs in the AI sector

09:28

also the absence of companies with no

09:30

intrinsic value going public and the

09:32

lack of quote funky AI companies is a

09:34

key indicator he noted that the current

09:36

market doesn't exhibit such

09:38

characteristics and I think that I

09:39

largely agree with that there's

09:41

obviously some bad players in the

09:43

industry who are just trying to make a

09:44

quick Buck but there might be more to it

09:47

and here's what I mean there's a known

09:48

pattern for emerging Technologies called

09:50

the Gartner hype cycle it describes how

09:53

humans tend to be overzealous about a

09:54

new revolutionary technology we

09:57

overestimate and overshoot its impact

09:59

hype it up until the market crashes and

10:02

after the crash the real companies with

10:04

the real value come through and

10:06

spearhead the development of the new

10:08

technology until it's mature if we look

10:10

at the Gardner hype cycle for AI it

10:12

seems like we're at the closing stages

10:14

of the peak of inflated expectations and

10:17

are slowly entering the trough of

10:18

disillusionment this interpretation of

10:21

the current market means that there's

10:22

going to be a lot of disappointment

10:23

ahead the real applications of AI will

10:26

come after the hype and fatigue Fades

10:29

but but this time could be a little bit

10:30

different because at its core AI has the

10:33

ability to mimic cognitive labor a

10:35

change that no previous technology has

10:37

been able to do without intrinsic human

10:40

intervention whereas as pointed out in a

10:42

previous podcast episode during the dot

10:45

bubble there were a lot of people simply

10:46

making a random website and going public

10:49

and investors would throw money at

10:50

anything with a do at the end of its

10:56

name there's been a lot of organizations

10:58

that have been ring to launch AI

11:00

products or replace humans with AI

11:02

altogether but the ones who are hastily

11:04

adding the ever popular quote powered by

11:06

AI gimmick are failing badly and there

11:09

have been some spectacular AI failures

11:11

for example a parcel delivery firm

11:14

called dpd enabled an AI online chat for

11:17

answering questions for One customer it

11:19

was not only useless at answering

11:21

relevant questions regarding his order

11:23

but instead when asked it provided him

11:25

with a poem and swore at him this

11:27

amazing poem came forth about a dpd

11:29

chatot who who wasn't able to do

11:32

anything it was quite amazing it

11:33

appeared to even disregard its own rules

11:36

and started swearing and even started

11:38

dissing its own employer adding dpd is

11:41

the worst delivery firm in the world I

11:43

would never recommend them to anyone in

11:46

Australia a group of academics used AI

11:49

generated material as part of their

11:50

complaint submission to the Australian

11:52

parliament against four of Australia's

11:54

biggest banks of course the AI made up a

11:57

whole bunch of information including SC

11:59

scandals that never existed and

12:00

implicating deoe when they had nothing

12:02

to do with the original complaint

12:04

someone on the team didn't read the

12:05

entire submission before it went to the

12:07

Australian government so it was a huge

12:08

embarrassment and the academics were

12:10

forced to apologize this is just one of

12:12

many examples and I'm going to include a

12:14

website that keeps a running tab of

12:16

these AI blunders but on the other hand

12:19

there are companies who are integrating

12:21

AI successfully and this is where we

12:24

come to the other kind of broken promise

12:26

when it comes to AI it's not CEO's over

12:28

promising ai's usefulness but under

12:30

promising instead in other words we were

12:33

told that AI would be our assistant and

12:35

create new jobs not replace us but what

12:38

if that's just not true not about

12:40

replacing the human in the loop in fact

12:42

it's about empowering the human like

12:45

it's an assistant it's an assistant this

12:46

clip from SAA is a bit hard to believe

12:48

now even though later he did mention the

12:50

flip side yes I'm sure a lot of us use

12:53

chat GPT to help get work done on

12:56

personal projects but AI eventually

12:58

creating more jobs than it would replace

13:00

maybe not so fast you may have noticed

13:03

that since 2022 there's been a lot of

13:06

tech related layoffs The Narrative is

13:08

that this is happening due to the over

13:10

hiring during the pandemic times and

13:12

that would make sense for 2022 and even

13:14

2023 but in 2024 we're still seeing Mass

13:18

layoffs now if companies were just

13:20

struggling to make a profit the

13:22

explanation would be as simple as the

13:23

common business cycle but strangely some

13:26

of the biggest companies in the world

13:27

are seeing record profits they're almost

13:29

cyclically shedding

13:31

workers according to reports tech

13:34

companies fired 165,000 people in 2022

13:38

260,000 in 2023 and so far the

13:41

projection is 270,000 in 2024 so if

13:45

companies are doing fine why are workers

13:47

being kicked out the door well marketing

13:49

professor and author Scott Galloway

13:51

thinks quote there's something else

13:54

going on end quote in his recent blog

13:56

post dissecting the impact of AI on

13:59

corporations he mentions quote AI is

14:02

playing a larger role in layoffs than

14:03

CEOs are willing to admit end quot the

14:06

normal way we think about Job

14:07

Replacements caused by AI is that Joe

14:10

blogs is working at his computer

14:12

analyzing company figures or something

14:13

like that and then suddenly the next day

14:15

he's replaced by an AI but the truth is

14:18

it's a bit more nuanced than that as

14:21

companies slowly get better at

14:23

understanding how AI can Aid in their

14:24

operations they'll create better teams

14:26

and structure with higher outcomes

14:29

basically instead of three people doing

14:31

the work of three you'll have one person

14:33

doing the work of five with AI the

14:36

problem is that none of these companies

14:38

are being sincere about it because if

14:40

you're a CEO and you've been hyping up

14:42

AI as the single greatest thing to

14:43

happen to the industry and then you

14:45

follow it up with a memo that you've

14:46

laid off 30% of your staff because of AI

14:49

the hypocrisy becomes clear but in 2024

14:52

it seems like the truth is slowly coming

14:54

out and there's hints everywhere IBM

14:56

made headlines when CEO arvind chrish

14:59

mentioned that they're planning to pause

15:00

hiring roles that could be replaced with

15:02

AI same with UPS where recently their

15:05

CEO Carol Tom acknowledged that AI is

15:08

playing a part in employee layoffs and

15:10

this came after UPS had the largest

15:12

layoff in its 116e history 12,000

15:16

workers in both cases spokes people

15:19

later clarified saying quote AI is not

15:22

replacing workers end quote which is

15:24

basically PR talk to save face in fact a

15:27

Shopify employee broke their

15:29

non-disclosure agreement to expose that

15:31

the company has been secretly firing

15:33

employees because of AI the shift is

15:35

even having an impact far outside the

15:37

tech industry TV and media producer

15:40

Tyler Perry was building an $800 million

15:43

Studio expansion but paused its

15:44

construction after he saw the results of

15:46

open aiora an AI video generator he made

15:50

the remark quote I am very very

15:53

concerned that in the near future a lot

15:54

of jobs are going to be lost end quote

16:00

in the last couple of years you probably

16:02

would have heard this phrase a bit AI

16:04

will not replace humans but humans who

16:06

utilize AI replace humans who don't

16:08

there's definitely some truth to it but

16:10

some jobs will be at risk regardless we

16:13

don't have to imagine this it is already

16:14

happening Mega caps like meta and

16:16

alphabet they are shrinking their

16:18

workforces and and shrinking their

16:20

hiring it's this idea that senior AI

16:22

researchers or data analysts they can do

16:24

more with Gen than an army of Junior

16:27

Engineers who code a recent report from

16:30

Mackenzie found that around 12 million

16:32

us workers may need to switch jobs by

16:34

2030 lower wage workers are 14 times

16:37

more likely to be impacted stem

16:40

Healthcare Builders and professional

16:42

Fields will continue to add jobs but

16:44

generative AI could change work

16:46

significantly for occupations like

16:48

customer service and sales and office

16:50

support there's been repeated reports of

16:52

how AI will be adding hundreds of

16:54

millions of jobs in the next decade and

16:56

how it's going to positively impact GDP

16:59

while by some miracle this could happen

17:01

right now as has been the case in many

17:03

times in history the train will leave

17:05

the station without telling us whether

17:06

we're aware to jump on board or not is

17:08

the

17:11

question so wrapping this all up I do

17:14

have to mention that in no way am I

17:16

saying that all forms of AI and all of

17:18

these giant corporations are elaborate

17:20

hoaxes in fact on this channel I've been

17:23

making videos about AI progress for the

17:25

better part of a decade now and I've

17:26

seen exponential progress right in front

17:28

of my eyes and so will many of you we

17:31

may have hit a temporary bottleneck when

17:32

it comes to compute but all it takes is

17:34

a breakthrough in model efficiency and

17:36

the curve continues upwards but I just

17:38

think that it's important to still have

17:39

a cautious approach when it comes to

17:41

looking at all sides of a story I'm also

17:43

working on a video that's all about the

17:45

positive real life impacts of AI so

17:47

hopefully that's going to balance things

17:49

out a

17:50

bit so far it seems like the promises of

17:52

AI are yet to be fully realized whether

17:55

that's creating and not taking jobs

17:57

solving Global issues or even providing

17:59

us with a Dependable AI customer service

18:01

spot that doesn't swear at us with such

18:04

big and bold promises from these

18:05

companies over the years it's important

18:07

for all of us to ask ourselves do we

18:09

have technological gold a trojan horse

18:12

or a chess master hiding under a

18:16

table anyways that's about it from me my

18:18

name is toogo and You' been watching

18:20

cold fusion if you did like that episode

18:22

feel free to subscribe there's plenty of

18:24

other interesting stuff on the Channel

18:26

all right so I'll catch you again soon

18:28

cheers guys have a good

18:37

one a better

18:39

future Heaven Help Us make a better

18:42

[Music]

18:55

future cold fusion it's new