What If AI Becomes Self Aware? | Unveiled
Category
😹
FunTranscript
00:00 In June of 1956, Dwight Eisenhower was President of the United States.
00:05 Interstate highways were first being built all across America, and Elvis was exploding
00:10 onto the scene with Heartbreak Hotel.
00:12 At the same time, on the campus of Dartmouth College, dozens of mathematicians and scientists
00:17 met for a six-week workshop.
00:20 They were there to discuss a topic so new and so esoteric, it did not yet have a name.
00:25 At this Dartmouth workshop, these top brains debated how to go about creating a thinking
00:30 machine.
00:31 The Dartmouth Summer Research Project on Artificial Intelligence gave birth to a new field of
00:36 science.
00:37 Less than a century later, that field is poised to reshape the world.
00:42 This is Unveiled, and today we're answering the extraordinary question; what if AI becomes
00:47 self-aware?
00:49 Do you need the big questions answered?
00:51 Are you constantly curious?
00:52 Then why not subscribe to Unveiled for more clips like this one?
00:55 And ring the bell for more thought-provoking content!
00:59 The advent of the microprocessor has led to a technological explosion.
01:04 As the size of computers has diminished, their processing power has increased.
01:08 The computers that took astronauts to the moon in 1969 filled an entire room.
01:13 Today, the smartphone in your pocket is millions of times more powerful.
01:17 That rapid evolution has paved the way for the wildest dreams of the Dartmouth workshop
01:22 to come true.
01:23 In 2023, individuals, businesses and governments around the world are reckoning with the integration
01:29 of AI into everyday life.
01:31 But before understanding the ifs, hows and whys of machine self-awareness, we must first
01:36 distinguish between forms of AI.
01:39 There are several types of artificial intelligence; generative AI, traditional AI and artificial
01:45 general intelligence (AGI).
01:48 Generative AI chatbots and art generators have exploded into pop culture since 2022
01:53 and are becoming a commonplace tool for businesses and individuals around the world.
01:57 A prompt is input into the AI, which creates content based on the structured data to which
02:02 it has access.
02:04 Users were initially delighted by the art and text created by generative AI systems…
02:08 but as time went on, there's also been some backlash, as artists in particular object
02:13 to how their art has been used to train machine learning models without compensation.
02:19 Generative AI is limited both by the flaws of the human who created it and the flaws
02:23 in the data it scrapes, which includes sources such as websites, books and social media.
02:28 Studies out of John Hopkins and the Georgia Institute of Technology, for example, have
02:32 found that virtual robots trained on images and captions from the internet can become
02:37 racist and sexist.
02:39 Contrast that with traditional artificial intelligence.
02:42 Artificial AI requires complex algorithms that specialise in pattern recognition, honing
02:47 analytics, fraud detection and data classification.
02:50 It's a less intelligent form of AI built to focus on and achieve specific tasks.
02:56 When we think of a machine coming alive, we're talking about AGI.
03:01 Artificial General Intelligence is an autonomous system that learns, evolves and ultimately
03:05 leapfrogs human ability.
03:07 There are multiple companies and researchers around the world who are seeking to create
03:10 AGI.
03:12 But the public and private sectors are falling over themselves trying to use AI to improve
03:16 operations.
03:17 The potential uses of AI in Hollywood was a central sticking point in the 2023 WGA and
03:23 SAG-AFTRA strikes, for example.
03:25 Google is testing AI in cardiac medicine.
03:28 Wall Street uses AI to get an edge in a stock market where millions can be made or lost
03:33 in microseconds.
03:34 American Express has started to use AI bots for customer service operations.
03:39 Microsoft has committed $50 million to its AI for Earth initiative, which provides tools
03:44 and offers grants in four key areas - agriculture, biodiversity, climate and water.
03:50 AIs of these types are surpassing the abilities of their human architects.
03:55 Self-improving, independent algorithms already exist.
03:58 Even researchers who were bullish on the capabilities of AI have been stunned by the rapidity of
04:03 its evolution.
04:04 But as society reconciles with what it means to use AI every day, we're hurtling towards
04:09 a time where AI is as ubiquitous as the smartphone.
04:13 That begs the question - what does a society suffused with AI look like?
04:19 Everyday life will likely involve various types of AIs interacting with one another,
04:23 each specializing in complex tasks, working together like cogs in a machine.
04:27 The first level of impact will be seen in the workforce.
04:31 Millions of human jobs will be supplanted by AI-guided robots.
04:34 In fact, the World Economic Forum estimates that AI will displace 85 million jobs by 2025,
04:41 although it will also create 97 million new jobs.
04:45 Media, medicine, finance, construction, logistics - there is no field that will remain untouched.
04:51 Some governments may utilize the increased economic output generated by AI to establish
04:56 a universal basic income.
04:58 In other countries, AI will add to the growing problem of wealth and income inequality.
05:03 It could also cause major disruptions to industries around the world, leading to political destabilization.
05:08 Still, it's difficult to argue against some of the benefits to society brought by artificial
05:13 intelligence.
05:14 Brown University researchers have been working on robots to help seniors keep track of their
05:18 medication schedules.
05:20 Climate scientists think that AI could help us mitigate and perhaps even reverse climate
05:24 change.
05:25 Some researchers envision a future in which swarms of AI drones reflect sunlight away
05:30 from our warming oceans.
05:31 There are also a myriad of applications for AI-run nanobots.
05:35 In the medical field, AI-run nanobots could turn people into cyborgs of a sort, repairing
05:41 our bodies and making humans effectively immortal.
05:44 Scientists are also finding applications in astrophysics and space exploration.
05:48 Swiss computer science expert, Jorgen Schmidhuber, at the Dalai Malay Institute for Artificial
05:54 Intelligence envisions a utopian future, predicting that "AIs will colonize and transform the
05:59 entire cosmos, and they will make it intelligent."
06:03 As AI progresses, it's possible that we'll see innovations coming at a pace so rapid
06:08 that it makes the technological revolutions of the 20th century appear quaint by comparison.
06:13 We could live in "smart cities", with municipal functions being run by predictive
06:17 AI.
06:18 Energy, waste management, public transportation and utilities may all be delegated to machines,
06:24 leading to net-zero carbon cities.
06:27 AIs could also revolutionise recycling, drastically reducing human-created waste.
06:32 As in Star Trek, humanity could be on its way to an AI-driven era of post-scarcity,
06:38 with AI-run machines altering material on the cellular or atomic level.
06:43 But, as we've discussed, artificial intelligence can also be limited by the biases of its creators.
06:48 Public ethicists are fast becoming some of the most important voices in public policy.
06:53 They and other AI researchers have been sounding the alarm bells about the unintended consequences
06:59 of the impending AI revolution.
07:01 They're pushing for the development of ethical AI, incorporating ethics and morality into
07:06 the systems as we build them.
07:08 A facial recognition AI program, for example, needs to be trained to not discriminate based
07:13 on race.
07:14 As more of the world is run on AI, and as AI evolves to the point of sentience, we need
07:19 to ensure that it's programmed without bias and with a respect for human life and dignity.
07:24 Otherwise, we could be looking at a Skynet scenario.
07:27 If a machine ever achieved sentience, odds are good that it'll be in artificial general
07:32 intelligence.
07:33 After that point, we could see what's called a "technological singularity" - an intelligence
07:38 explosion - in which an AI self-improves in an escalating cycle, until it far surpasses
07:43 human capabilities.
07:45 Faced with such a superintelligence, we'd be unable to control or reverse it.
07:49 This is a possibility that several notable scientists, including Stephen Hawking, have
07:54 expressed serious concern about.
07:56 We don't know for sure, however, that the growth of such an AI's knowledge would be
07:59 exponential.
08:01 There may be natural limits to what AI can know and do.
08:04 For example, some scientists think that AI could be hampered by current limitations in
08:09 technology and that exponential increase in intelligence could end up leading to diminishing
08:14 returns.
08:15 Still, what we can be sure of is that AI will be disruptive.
08:19 While tech bros are fond of the phrase "move fast and break things", disruption can have
08:24 plenty of negative, knock-on effects.
08:26 Even if a superintelligent AI is benevolent, the disruption it could cause would be immense.
08:31 Virtually every industry on Earth could be remade from the ground up.
08:35 If human workers were replaced en masse, we would need to consider a universal basic income.
08:41 Even then, without the need to work, many would begin to question their purpose in the
08:45 world.
08:46 Depending on the rapidity of AI's rise, anarchy could spread around the globe.
08:51 And if AI's become self-aware, the ethical and moral questions multiply.
08:56 Should intelligent, self-aware AI's be considered alive?
08:59 Should we view them as having individual personhood?
09:02 What moral principles bind them, and what laws apply?
09:05 Countries, companies and the international community would need to debate the question
09:09 - do machines have human rights?
09:11 After all, the word "robot" is derived from a Czech word meaning "slave".
09:16 Humans are affected both by nature and nurture.
09:18 Ideally, our self-awareness aids us in exercising self-control.
09:23 If sentient machines are inevitable, as some researchers believe, it's imperative that
09:27 we worry about ethics now, while they're in their infancy.
09:31 For decades, science fiction writers have feared the overthrow of humanity by our own
09:35 robotic creations.
09:37 While anything is possible, such a dystopian future becomes less likely if we raise our
09:42 mechanical children with ethics and morality in mind.
09:45 AI will be disruptive, no question.
09:47 But as technology rapidly strides towards sentience, we have opportunities and serious
09:52 issues to answer.
09:54 Will what we create benefit humanity, or lead to more suffering?
09:58 Will our creations be beholden to us as slaves, or partners who can help free us from our
10:03 own limitations?
10:04 All are possibilities, and all could be what happens if AI becomes self-aware.
10:11 What do you think?
10:12 Is there anything we missed?
10:13 Let us know in the comments, check out these other clips from Unveiled, and make sure you
10:17 subscribe and ring the bell for our latest content.
10:20 content.