Why AI progress seems "stuck" | Jennifer Golbeck | TEDxMidAtlantic

Summary

The video script discusses the current state of artificial intelligence, highlighting the difference between narrow AI and the hypothetical AGI. It questions the hype around AGI, suggesting that current AI tools, while powerful, are not yet at a level that poses an existential threat. The speaker addresses concerns about AI's reliability and potential for 'hallucinations,' the challenges in improving AI with data, and the economic feasibility of advancements. They also touch on the impact of AI on jobs, the issue of AI bias, and the unique aspects of human intelligence that AI cannot replicate, concluding with a humorous note on humanity's control over AI.

Takeaways

• 🧠 The script discusses the current state of artificial intelligence (AI), highlighting that while AI has surpassed human performance in specific tasks, such as chess, the concept of artificial general intelligence (AGI) is still a topic of debate and concern.
• 🔮 There is a growing concern among some in the tech industry about the potential dangers of AI, with calls for regulation due to its perceived power and threat to civilization.
• 💡 The script suggests two reasons for the industry's focus on the dangers of AI: the potential for significant financial gain by emphasizing the power of the technology, and the cinematic allure of AI overtaking humanity, which distracts from current AI-related issues.
• 🤖 The speaker questions the likelihood of achieving AGI, citing examples of current AI tools that are not yet perfect, such as Google's AI Search tool and the unreliability of generative AI in producing accurate responses.
• 📈 The script points out that while there is significant investment in AI, the return on investment is not yet clear, and the sustainability of this investment is in question.
• 🔍 A key challenge identified for AI is reliability, with algorithms often providing incorrect information, which is a hurdle to overcome before AI can live up to its hype.
• 🎭 The concept of 'AI hallucination' is introduced, referring to the tendency of AI to fabricate information or responses, which is a significant issue that needs addressing.
• 👨‍💼 The script argues that the fear of AI taking jobs may be overstated, as increased efficiency through AI tools could lead to profit rather than job losses.
• 🔒 The issue of AI inheriting human biases is highlighted as a persistent problem that has not been solved, which is crucial when considering AI in decision-making roles.
• 🛠️ The speaker expresses skepticism about the ability to solve the reliability and hallucination problems of AI, suggesting that the technology may have reached a plateau.
• 🌐 Finally, the script emphasizes that human intelligence is defined by our emotional and creative capabilities, which AI cannot replicate, offering a reassurance that AI will not replace our core humanity.

Q & A

What is the concept of Artificial General Intelligence (AGI) discussed in the script?

AGI refers to the idea of AI that can perform at or above human levels on a wide variety of tasks, similar to the capabilities of human intelligence.

Why are some people in the tech industry concerned about the AI they are building?

They believe the AI is so powerful and dangerous that it poses a threat to civilization and may need to be regulated due to its potential to cause existential harm to humanity.

What are the two main reasons suggested in the script for the tech industry's concern about AI?

One is the potential for significant financial gain by emphasizing the power and danger of their technology, and the other is the cinematic appeal of the concept of AI overtaking humanity, which serves as a distraction from real-world AI problems.

What is the current state of AI in terms of achieving AGI according to the script?

The script suggests that while there is hype around AGI, the current state of AI, exemplified by tools like Google's AI Search, is far from achieving AGI and may be at a plateau rather than on a sharp upward trajectory.

What is the main challenge that needs to be solved to realize the hype around AI?

The main challenge is reliability, as AI algorithms often produce incorrect results or 'hallucinations,' which means they cannot be fully trusted to perform tasks without human correction.

What is an 'AI hallucination' as mentioned in the script?

AI hallucination refers to the phenomenon where AI makes up information or content that did not exist in the training data, leading to incorrect or misleading outputs.

Why is solving the AI hallucination problem important for the future of AI?

Solving the hallucination problem is important because it affects the reliability and trustworthiness of AI, which are crucial for AI to live up to its hype and be useful in practical applications.

What are the two factors mentioned in the script that AI tools need to improve upon?

AI tools need to improve upon the amount of data they are trained on and the underlying technology itself to enhance their capabilities and reliability.

How does the script address the concern about AI taking all of our jobs?

The script suggests that the concern is based on a misunderstanding, as AI can increase efficiency but does not necessarily replace jobs, especially considering the cost and availability of AI tools.

What is the fundamental issue with AI that the script suggests we should worry about?

The script suggests worrying about the issue of AI adopting human biases from training data, which has not been successfully addressed and can lead to problematic outcomes in decision-making.

What is the final point made in the script about human intelligence and AI?

The script concludes that human intelligence is defined by our ability to connect, have emotional responses, and creatively integrate information, which AI cannot replicate, thus distinguishing our humanity from AI capabilities.

Outlines

00:00

🧠 The Hype and Concerns Around AGI

The first paragraph discusses the current state of artificial intelligence (AI), highlighting its ability to outperform humans in specific tasks such as playing chess. It introduces the concept of artificial general intelligence (AGI), which is AI that can perform at or above human levels across a wide range of tasks. The speaker expresses concern about the discourse surrounding AGI, noting that while some in the tech industry warn of its potential dangers to civilization, others may be motivated by financial gain or the cinematic allure of AI overtaking humanity. The paragraph also points out that focusing on improbable futures can distract from real-world issues already arising from AI, such as racial bias in AI decision-making for prison release and the challenge of deep fakes.

05:01

🔮 The Reality of AGI and the Challenge of Reliability

The second paragraph delves into the challenges of achieving AGI, starting with the issue of reliability. It mentions AI's tendency to produce incorrect results, using Google's AI Search tool as an example. The speaker argues that the current trajectory of AI improvements may not be sufficient for achieving AGI and discusses 'AI hallucination,' where AI generates false information or images. The paragraph also addresses the high expectations set for AI in fields like law, where it has been used to write legal briefs, only to generate fictitious cases. The need for more data and technological advancements is highlighted, along with skepticism about the availability of sufficient high-quality data and the sustainability of investment in AI improvement.

10:01

🛠 The Future of AI: Improvements, Bias, and Human Connection

The final paragraph contemplates the future of AI, focusing on the need for substantial improvements in data and technology. It questions the economic viability of investing in AI to replace human workforces, given the availability of affordable, open-source AI tools. The speaker emphasizes the persistent issue of AI inheriting human biases and the futility of guardrails in addressing this problem. The paragraph concludes by distinguishing between human intelligence, defined by emotional connection and creativity, and AI, which lacks these core human attributes. It reassures that despite fears of AI overlords, humans retain control over technology, as we can always 'turn it off.'

Keywords

💡Artificial Intelligence (AI)

Artificial Intelligence refers to the simulation of human intelligence in machines that are programmed to think like humans and mimic their actions. In the video, AI is central to the discussion, with a focus on its current capabilities and potential future developments, especially in relation to surpassing human performance in various tasks.

💡Generative AI

Generative AI is a subset of AI that can generate new content, such as text, images, or music, rather than just recognizing patterns or making predictions. The video script discusses the introduction of generative AI to the public and the ensuing discussions about its potential to achieve artificial general intelligence (AGI).

💡Artificial General Intelligence (AGI)

AGI represents the concept of AI that can perform at or above human levels across a broad range of tasks, as opposed to narrow AI that excels in specific tasks. The script explores the idea of AGI and the concerns surrounding the potential for AI to match or exceed human cognitive abilities.

💡Reliability

Reliability, in the context of AI, pertains to the consistency and accuracy of the AI's performance. The video emphasizes the importance of reliability in AI, noting that current algorithms often produce incorrect results, which undermines their usefulness and trustworthiness.

💡AI Hallucination

AI hallucination is a term used to describe the phenomenon where AI generates content that is factually incorrect or makes things up, even when presented with real data. The script provides examples of AI hallucination, such as creating non-existent threats of violence or generating misleading legal cases.

💡Deep Fakes

Deep fakes are AI-generated media, often videos or audio, that convincingly replace or superimpose people's images or voices without their consent. The video mentions deep fakes as an example of real-world problems caused by AI that deserve attention instead of focusing on speculative AGI scenarios.

💡Racial Bias

Racial bias in AI refers to the不公平 treatment or discrimination against certain racial groups by AI systems, often due to biased training data. The script discusses the issue of racial bias in AI, particularly in the context of AI being used in decision-making processes such as parole board decisions.

💡Incremental Improvements

Incremental improvements suggest small, gradual advancements in a technology or system. The video contemplates whether the current trajectory of AI development is leading towards AGI or if it is simply reaching a plateau with only incremental improvements expected in the future.

💡Productivity

Productivity in the video is discussed in the context of AI's potential to enhance human efficiency in the workplace. However, it also challenges the notion that increased productivity is the sole measure of human intelligence, emphasizing the importance of emotional intelligence and creativity.

💡Human Intelligence

Human intelligence encompasses not just cognitive abilities but also emotional responses, social connections, and creative thinking. The video contrasts human intelligence with AI, asserting that AI may imitate but cannot truly replicate the depth and complexity of human intelligence.

💡Elon Musk

Elon Musk, mentioned in the script, is a notable figure in the tech industry known for his views on AI, including predictions about the timeline for achieving AGI. His perspective adds to the debate on the pace of AI development and its potential impact on society.

💡Google AI Search

Google AI Search is an example of a specific AI application discussed in the video, which aims to provide direct answers to users' queries. The script uses it to illustrate the current limitations of AI in providing reliable and accurate information.

Highlights

Artificial intelligence has surpassed human performance in specific tasks such as chess.

The concept of artificial general intelligence (AGI) is gaining attention, with concerns about its potential threat to civilization.

Tech industry leaders are warning about the dangers of AI, advocating for regulation.

There is skepticism about the profitability and necessity of regulating powerful AI tools.

The fear of AI overtaking humanity is often sensationalized, distracting from current AI-related issues.

Elon Musk predicts AGI could be achieved within a year, despite current AI tools' limitations.

Google's AI Search tool exemplifies the current limitations of AI in providing accurate information.

The trajectory of AI development needs to be continuously upward to achieve AGI.

Reliability is a significant challenge for AI, as algorithms often produce incorrect results.

AI 'hallucination', or making up information, is a major issue that needs addressing.

The potential solution to AI hallucination may not be achievable with current technology.

Legal applications of AI have faced issues with accuracy and the creation of fictitious cases.

Even the best AI tools still hallucinate a significant percentage of the time.

The need for more data and technological improvements to enhance AI capabilities.

The challenge of finding reliable data to train AI, especially with the prevalence of low-quality content.

Investment in generative AI has not yet resulted in a sustainable financial return.

The debate over AI replacing jobs and the economic implications of increased efficiency.

AI's inability to replicate human emotional intelligence and creativity.

The persistent issue of AI inheriting human biases and the challenges in addressing this.

The importance of solving AI bias before widespread adoption in decision-making roles.

A reminder that, contrary to movies, we can always turn off AI if it becomes a threat.

Transcripts

[Music]

we've built artificial intelligence

already that on specific tasks performs

better than humans there is AI that can

play chess and beat human Grand Masters

but since the introduction of generative

AI to the general public a couple years

ago there's been more talk about

artificial general intelligence or AGI

and that describes the idea that there's

AI that can perform at or above human

levels on a wide variety of tasks just

like we humans are able to do and people

who think about AGI are worried about

what it means if we reach that level of

performance in the technology right now

there's people from the tech industry

coming out and saying the AI that we're

building is so powerful and dangerous

that it poses a threat to civilization

and they're going to government and

saying maybe you need to regulate us now

normally when an industry makes a

powerful new tool they don't say it

poses an existential threat to humanity

and that it needs to be limited so why

are we hearing that language and I think

there's two main reasons one is if your

technology is so powerful that it can

destroy civilization between between now

and then there's an awful lot of money

to be made with that and what better way

to convince your investors to put some

money with you than to warn that your

tool is that dangerous the other is that

the idea of AI overtaking humanity is

truly a cinematic concept we've all seen

those movies and it's kind of

entertaining to think about what that

would mean now with tools that we're

actually able to put our hands on in

fact it's so entertaining that it's a

very effective distra action from the

real problems already happening in the

world because of AI the more we think

about these improbable Futures the less

time we spend thinking about how do we

correct deep fakes or the fact that

there's AI right now being used to

decide whether or not people are let out

of prison and we know it's racially

biased but are we anywhere close to

actually achieving AGI some people think

so Elon Musk said that we'll achieve it

within a year I think he posted this a

few weeks ago but like at the same time

Google put out their AI Search tool

that's supposed to give you the answer

so you don't have to click on a link and

it's not going super

well please don't eat

rocks now of course these tools are

going to get better but if we're going

to achieve AGI or if they're even going

to fundamentally change the way we work

we need to be in a place where they are

continuing on a sharp upward trajectory

in terms of their abilities and that may

be one path but there's also the

possibility that what we're seeing is

that these tools have basically achieved

what they're capable of doing and the

future is incremental improvements in a

plateau so to understand the AI future

we need to look at all the hype around

it and get under there and see what's

technically possible and we also need to

think about where are the areas that we

need to worry and where are the areas

that we don't so if we want to realize

the hype around AI the one main

challenge that we have to solve is

reliability these algorithms are wrong

all the time like we saw with Google and

Google actually came out and said after

these bad search results um were

popularized that they don't know how to

fix this problem I use chat GPT every

day I read a newsletter that summarizes

discussions on farri message boards and

so I download that data chat GPT helps

me write a summary and it makes me much

more efficient than if I had to do it by

by hand

but I have to correct it every day

because it misunderstands something it

takes out the context and so because of

that we can't just rely on it to do the

job for me and this reliability is

really important now a subpart of

reliability in this space is AI

hallucination a great technical term for

the fact that AI just makes stuff up a

lot of the time I did this in my

newsletter I said J gbt are there any

people threatening violence if so give

me the quotes and It produced these

three really clear threats of violence

that didn't sound anything like people

talk on these message boards and I went

back to the data and nobody ever said it

it just made it up out of thin air and

you may have seen this if you've used an

AI image generator I asked it to give me

a close up of people holding hands

that's a hallucination and a disturbing

one at

that we have to solve this hallucination

problem if this AI is going to live up

to the hype and I I don't think it's a

solvable problem with the way this

technology works there are people who

say we're going to have it taken care of

in a few months but there's no technical

reason to think that's the case because

generative AI always makes stuff up when

you ask it a question it's creating that

answer or creating that image from

scratch when you ask it's not like a

search engine that goes and finds the

right answer on a page and so because

its job is to make things up every time

I don't know that we're going to be able

to get it to make up correct stuff and

then not make up other stuff that's not

what it's trained to do and we're very

far from achieving that and in fact

there are spaces where they're trying

really hard one space that there's a lot

of enthusiasm for AI is in the legal

area where they hope it will help write

legal briefs or do research some people

have found out the hard way that they

should not write legal briefs right now

with chat GPT and send them to federal

court because it just makes up cases

that sound right and that's a real

really fast way to get a judge mad at

you and to get your case thrown out now

there are legal research companies right

now that advertise hallucination free

generative Ai and I was really dubious

about this and researchers at Stanford

actually went in and checked it and they

found the best performing of these

hallucination free tools still

hallucinate 177% of the time so like on

one hand it's a great scientific

achievement that we have built a tool

that we can posee basically any query to

and 60 or 70 or maybe even 80% of the

time it gives us a reasonable answer but

if we're going to rely on using those

tools and they're wrong 20 or 30% of the

time there's no model where that's

really useful and that kind of leads us

into how do we make these tools that

useful because even if you don't believe

me and you think we're going to solve

this hallucination problem we're going

to solve the reliability problem the

tools still need to get better than they

are now and there's two things they need

to do that one is lots more data and two

is the technology itself has to improve

so where are we going to get that data

because they've kind of taken all the

reliable stuff online already and if we

were to find twice as much data as

they've already had that doesn't mean

they're going to be twice as

smart I don't know if there's enough

data out there and it's compounded by

the fact that one way that generative AI

has been very successful is at producing

lowquality content online that's bots on

social media misinformation and these

SEO pages that don't really say anything

but have a lot of ads and come up high

in the search results and if the AI

starts training on pages that it

generated we know from Decades of AI

research that they just get

progressively worse it's like the

digital version of mad cow

disease let's say we solve the data

problem you still have to get the

technology better and we've seen $50

billion dollar in the last couple years

invested in improving generative Ai and

that's resulted in $3 billion in Revenue

so that's not sustainable but of course

it's early right companies may find ways

to start using this technology but is it

going to be valuable enough to justify

the tens and maybe hundreds of billions

of dollars of Hardware that needs to be

bought to make these models get better I

don't think so and we can kind of start

looking at practical examples to figure

that out and it leads us to think about

where are the spaces we need to worry

and not because one place that

everybody's worried with this is that AI

is going to take all of our jobs lots of

people are telling us that's going to

happen and people are worried about it

and I think there's a fundamental

misunderstanding at the heart of that so

imagine this scenario we have a company

and they can afford to employ two

software engineers and if we were to

give those software Engineers some

generative AI to help write code which

is something it's pretty good at let's

say they're twice as efficient that's a

big overestimate but it makes the math

easy easy so in that case the company

has two choices they could fire one of

those software Engineers because the

other one can do the work of two people

now or they already could afford two of

them and now they're twice as efficient

so they're bringing in more money so why

not keep both of them and take that

extra profit the only way this math

fails is if the AI is so expensive that

it's not worth it but that would be like

the AI is $100,000 a year to do one

person's work of work so that sounds

really expensive and practically there

are already open-source versions of

these tools that are low cost that

companies can install and run themselves

now they don't perform as well as the

flagship models but if their half is

good and really cheap wouldn't you take

those over the one that cost a $100,000

a year to do one person's work of course

you would and so even if we solve

reliability we solve the data problem we

make the models better the fact that

there are cheap versions of this

available suggest that companies aren't

going to be spending hundreds of

millions of dollars to replace their

Workforce with AI there are areas that

we need to worry though because if we

look at AI now there are lots of

problems that we haven't been able to

solve I have been building artificial

intelligence for over 20 years and one

thing we know is that if we train AI on

human data the AI adopts human biases

and we have not been able to fix that

we've seen those biases start showing up

in generative AI

and the gut reaction is always well

let's just put in some guard rails to

stop the AI from doing the biased thing

but one that never fixes the bias

because the AI finds a way around it and

two the guard rails themselves can cause

problems so Google had an has an AI

image generator and they tried to put

guardrails in place to stop the bias in

the results and it turned out it made it

wrong this is a request for a picture of

the signing of the Declaration of

Independence and it's a great picture

but it is not factual correct and so in

trying to stop the bias we end up

creating more reliability problems we

haven't been able to solve this problem

of bias and if we're thinking about

deferring decision-making replacing

human decision makers and relying on

this technology and we can't solve this

problem that's a thing that we should

worry about and demand solutions to

before it's just widely adopted and

employed because it's sexy and I think

there's one final thing that's missing

here which is our human intelligence is

not defined by our productivity at work

at its core it's defined by our ability

to connect with other people our ability

to have emotional responses to take our

past and integrate it with new

information and creatively come up with

new things and that's something that

artificial intelligence is not now nor

will it ever be capable of doing it may

be able to imitate it and give us a

cheap fact simile of genuine connection

and empathy and creativity but it can't

do those core things to our humanity and

that's why I'm not really worried about

AGI taking over civilization but if you

come away from this disbelieving

everything I have told you and right now

you're worried about Humanity being

destroyed by AI overlords the one thing

to remember is despite what the movies

have told you if it gets really bad we

still can always just turn it

off thank you

MEC