Fall 2021
Dear Friends and Family,
I inserted this segment to give you some basics of AI technology, so you have some semblance of what it is, how it works, why more and more data is so important, and to get a feel for where the technology is headed.
My Takeaways:
Next: The series has alluded to Issues and Concerns with AI. This segment will look at such issues and concerns in some depth. Additionally, a potential timeline for AI is included.
Happy Learning,
Harley
Dear Friends and Family,
I inserted this segment to give you some basics of AI technology, so you have some semblance of what it is, how it works, why more and more data is so important, and to get a feel for where the technology is headed.
My Takeaways:
- Neural networks are a basic aspect of AI. It is a system that replicates the human brain and as such has the capacity for learning. Its learning is enhanced by being introduce to more data.
- Deep Learning was a huge breakthrough in the mid-2000s, when researchers’ discovered a way to train additional new layers in neural networks. This discovery multiplied the learning capability of machines by giving them the power of speech and object recognition.
- Machine learning has advanced to the point of emulating human emotion, intuition, and mimicking human thought. The gap between human and machine learning is shrinking. Clearly, machine thinking is faster, and their memory is far more accurate.
- There are four waves to AI. The first two are already present and in use. The fourth, autonomous AI, is the one that will revolutionize our lives.
Next: The series has alluded to Issues and Concerns with AI. This segment will look at such issues and concerns in some depth. Additionally, a potential timeline for AI is included.
Happy Learning,
Harley
BIG TECH & AI – SEGMENT 8
UNDERSTANDING ARTIFICIAL INTELLIGENCE – EXCERPTS
NEURAL NETWORKS and DEEP LEARNING: In 1983 the field of AI had forked into two camps: the “rule-based” approach and the “neural networks” approach. Researchers in the rule-based camp attempted to teach computers to think by encoding a series of logical rules: If X, then Y. This approach worked well for simple and well-defined games but fell apart when the universe of possible choices or moves expanded. The “neural networks” camp, however, took a different approach. Instead of trying to teach the computer the rules that had been mastered by a human brain, these practitioners tried to reconstruct the human brain itself. Neural networks require large amounts of two things: computing power and data. The data “trains” the program to recognize patterns by giving it many examples, and the computing power lets the program parse those examples at high speeds. The internet has led to an explosion of all kinds of digital data: text, images, videos, clicks, purchases, Tweets, and so on. Taken together all of this has given researchers copious amounts of rich data on which to train their networks, as well as plenty of cheap computing power for that training.
But the networks themselves were still severely limited in what they could do. Accurate results to complex problems required many layers of artificial neurons, but researchers hadn’t found a way to efficiently train those layers as they were added. Deep learning’s big technical breakthrough finally arrived in the mid-2000s when a leading researchers discovered a way to efficiently train those new layers in neural networks. The result was like giving steroids to the old neural networks, multiplying their power to perform tasks such as speech and object recognition. Deep learning brings AI’s power to bear on a range of real-world problems – the massive field to decipher human speech, translate documents, recognize images, predict consumer behavior, identify fraud, make lending decisions, help robots “see”, and even drive a car. Doing this requires massive amounts of relevant data, a strong algorithm, a narrow domain, and a concrete goal. Much of the difficult but abstract work of AI research has been done, and it is now time for entrepreneurs to roll up their sleeves and get down to the dirty work of turning algorithms into sustainable businesses.
AI ALGORITHMS: Today, successful algorithms need three things: big data, computing power, and the work of strong – but not necessarily elite – AI algorithm engineers. Bringing the power of deep learning to bear on new problems requires all three, but in this age of implementation, data is core. In deep learning, there’s no data like more data. The more examples of a given phenomenon a network is exposed to, the more accurately it can pick out patterns and identify things in the real world. Deep learning can only optimize what it can “see” by way of data.
Source: AI Superpowers by Kai-Fu Lee (2021)
MACHINE LEARNING: As machine learning continues to advance and expand the boundaries of AI, its algorithms become better at emulating human emotions and intuition. Models are continuously learning and adapting, resulting in machines that eerily mimic human thought as well as creative and emotional responses. The gap between human and machine intelligence is shrinking and intelligent machines often exceed human characteristics – in speed, accuracy, durability, cost efficiency, and endurance.
Traditional machine learning can function with relatively small amounts of data, but the performance of the algorithms tends to plateau after a certain amount of data has been assimilated. The creation of neural networks brought about a fundamental breakthrough in machine learning. This enabled the approach called deep learning that involves training neural networks on vast amounts of data and building practical applications. Deep learning algorithms keep improving as more data is fed into the system. Deep learning produces a virtuous cycle of success leading to more success. More data leads to superior products, which brings more users, who produce even more data, which increases profits, which further expands the business model and allows companies to collect and exploit even more data. The gap between the winners and losers widens and become almost impossible to close. A few well-known examples illustrate this:
DEVELOPMENT STAGES AND APPROACHES – A SCENARIO: Power Grids versus AI batteries: The giants aren’t just competing against one another in a race for the next deep learning. They’re also in a more immediate race against the small AI startups that want to use machine learning to revolutionize specific industries. It’s a contest between two approaches to distribute the “electricity” of AI across the economy: the “grid” approach of the seven giants versus the “battery” approach the of startups. How that race plays out will determine the nature of the AI business landscape – monopoly, oligopoly, or free-wheeling competition among hundreds of companies.
The “grid” approach is trying to commoditize AI. It aims to turn the power of machine learning into a standardized service that can be purchased by any company – or even be given away for free for academic or personal use – and accessed via cloud computing platforms. In this model, cloud platforms act as the grid, performing complex machines learning optimizations on whatever data problems users require. The companies behind these platforms – Google, Alibaba, and Amazon – act as the utility companies managing the grid and collecting the fees. AI startups are taking the opposite approach. Instead of waiting for this grid to take shape, startups are building highly specific “battery powered” AI products for each use. They build new products and train algorithms for specific tasks, including medical diagnosis, mortgage lending, and autonomous drones.
The Four Waves of AI: The complete AI revolution will take a little time and will ultimately wash over us in a series of four waves: Internet AI, business AI, perception AI and autonomous AI. The first two waves are already around us.
Differences in Approaching Global Markets: Silicon Valley giants want to directly introduce their products to global markets. They’ll make limited efforts to localization but will largely stick to the traditional playbook. They will build one global product and push it on billions of different users around the globe. Chinese companies are instead steering clear of direct competition and investing in the scrappy local startups that Silicon Valley looks to wipe out. It’s an approach rooted in the country’s own native experience. So, instead of seeking to both squash those startups and outcompete Silicon Valley, they’re throwing their lot in with the local businesses.
Source: AI Superpowers by Kai-Fu Lee (2018).
EVOLUTION OF AI DEVICES: In the decade spanning 2005 through 2015, the growth of smartphones positioned them for large-scale data collection. These devices, equipped with ever more advanced sensors, wireless networking, imaging, and voice capabilities, have been capturing much larger quantities of data than desktop or laptop counterparts. Social media companies grabbed this opportunity and became dominant platforms supplying a variety of services. Then the wearables space took off in a big way with several fitness-tracking watches quickly emerging with their functions eventually absorbed into smartwatches. The development of these devices established a new kind of data to mine at scale: health and fitness. In 2014, Amazon’s Alexa successfully overcame the first barrier to household penetration of smart home technologies. Amazon, Google, and Apple have made huge investments in home automation technologies and with appliance manufacturers to capture a plethora of data from households, including voice, video, temperature, soil humidity, air purity, Wi-Fi strength, how often someone opens the refrigerator, eating habits, and every other imaginable measurable by a gadget that’s connected to the AI hub at home. These voice interfaces will soon be embedded into ambient surfaces like ceiling and walls and furniture instead of being standalone devices.
Data privacy issues have not been adequately addressed vis-a-vis transparency on where and how all this data is stored and handled, who has access to it, and the relationship between the manufacturer and local law enforcement to govern its usage. From wearables, the industry is progressing to implants that physically fuse technology into our bodies and tinker with the creation of a new kind of augmented human being. These technologies may monitor our physiology and emotional responses in new ways and do so very accurately because of their proximity to the source of emotional responses in our bodies.
Each of the modalities [web browsers, smartphones, smart homes, wearables, implants] has broken new ground in creating a new kind of data about individuals that hitherto was unavailable or difficult to obtain. The more kinds of data that corporations can mine, the more complete the psychological and behavioral profiles of individuals can be built; hence the more sophisticated the ability of corporations to manipulate individuals and entire societies, by providing subtle nudges and cues through their every interaction with technology. Devices classified as wearables, implants, and smart homes are fundamentally changing not only how humans interact with machines, but ultimately influence the way humans think, feel, and behave.
Source: Artificial Intelligence and the Future of Power by Rativ Malhotra (2021)
CEDING CONTROL TO MACHINES: As AI systems advance and more of everyday life gets automated, the less control we actually have over the decisions being made about and for us. The current development track of AI prioritizes automation and efficiency which necessarily means we have less control and choice over the thousands of our everyday activities, even those that are seemingly insignificant. Our future living with AI begins with a loss of control over the little things. We’re not heading toward a single catastrophe but rather the steady erosion of the humanity we take for granted today. It’s time to see what happens as we transition from artificial narrow intelligence to artificial general intelligence – and what life will look like during the next 50 years as humanity cedes control to thinking machines.
Source: The Big Nine: How the Tech Titans & Their Thinking Machines Could Warp Humanity by Amy Webb (2019)
AI models and algorithms are touted as being objective and neutral, which is far from the truth. Biases of the people that control Ai systems enter at least two ways: through the data sets used for training and validation and via the precise definition of goals and objectives required for training machines. Machine learning models routinely make arbitrary decisions that the companies rationalize as an enforcement of community standards and community guidelines. But what are these standards and guidelines, who defines them, and how are they interpreted for enforcement? Social media companies have publicly established generic criteria like authenticity, misrepresentation, false information, respect for intellectual property, insensitivity to the dignity of others, hate speech, and so forth. These lofty principles are too broad and give the algorithms too much discretion. Though commendable at face value, in practice the implementation is inconsistent, biased, and in many cases, outright unfair. Worst of all, their algorithms lack transparency because the models that drive them are closely guarded secrets. Many posts have been rejected on spurious grounds that expose the digital platform’s cultural, racial, or ideological prejudice.
Source: Artificial Intelligence and the Future of Power by Rajiv Malhotra (2021)
USE OF DATA WITHOUT BREACH OF PRIVACY: Simply dissociating customers data from actual identities give insufficient protection and merely provides better optics in the eyes of regulators who do not know better. Using private data to make personalized shopping offers seems benign. However, the same advantage is afforded to a political group that seeks to influence votes, garner support for a cause, mobilize a revolt, or predict an individual’s behavior in a given negotiation or dispute. Such profiling of individuals is being used to develop psychological and emotional models likely to succeed, identify what threats and fears make the most effective deterrent, and enable other manipulative opportunities. These powerful impacts will pose serious challenges to social and political systems. But for now, AI is being hailed as a platform for utopia. The views of many futurists assume that society will proceed directly for the present social order to a superior one in the future. However, I foresee a prolonged period of disequilibrium in between, starting in the next decade and lasting until the end of this century. The youth of today will feel the brunt of this transition. By the time those born in 2020 are in their teens, the crisis of the present equilibrium will become obvious. And except for the lucky, sheltered elites that enjoy positions of power, most people will face uncertainty and turmoil.
Source: Artificial Intelligence and the Future of Power by Rajiv Malhotra (2021)
The unabbreviated version of the above can be found in the pdf document below.
UNDERSTANDING ARTIFICIAL INTELLIGENCE – EXCERPTS
NEURAL NETWORKS and DEEP LEARNING: In 1983 the field of AI had forked into two camps: the “rule-based” approach and the “neural networks” approach. Researchers in the rule-based camp attempted to teach computers to think by encoding a series of logical rules: If X, then Y. This approach worked well for simple and well-defined games but fell apart when the universe of possible choices or moves expanded. The “neural networks” camp, however, took a different approach. Instead of trying to teach the computer the rules that had been mastered by a human brain, these practitioners tried to reconstruct the human brain itself. Neural networks require large amounts of two things: computing power and data. The data “trains” the program to recognize patterns by giving it many examples, and the computing power lets the program parse those examples at high speeds. The internet has led to an explosion of all kinds of digital data: text, images, videos, clicks, purchases, Tweets, and so on. Taken together all of this has given researchers copious amounts of rich data on which to train their networks, as well as plenty of cheap computing power for that training.
But the networks themselves were still severely limited in what they could do. Accurate results to complex problems required many layers of artificial neurons, but researchers hadn’t found a way to efficiently train those layers as they were added. Deep learning’s big technical breakthrough finally arrived in the mid-2000s when a leading researchers discovered a way to efficiently train those new layers in neural networks. The result was like giving steroids to the old neural networks, multiplying their power to perform tasks such as speech and object recognition. Deep learning brings AI’s power to bear on a range of real-world problems – the massive field to decipher human speech, translate documents, recognize images, predict consumer behavior, identify fraud, make lending decisions, help robots “see”, and even drive a car. Doing this requires massive amounts of relevant data, a strong algorithm, a narrow domain, and a concrete goal. Much of the difficult but abstract work of AI research has been done, and it is now time for entrepreneurs to roll up their sleeves and get down to the dirty work of turning algorithms into sustainable businesses.
AI ALGORITHMS: Today, successful algorithms need three things: big data, computing power, and the work of strong – but not necessarily elite – AI algorithm engineers. Bringing the power of deep learning to bear on new problems requires all three, but in this age of implementation, data is core. In deep learning, there’s no data like more data. The more examples of a given phenomenon a network is exposed to, the more accurately it can pick out patterns and identify things in the real world. Deep learning can only optimize what it can “see” by way of data.
Source: AI Superpowers by Kai-Fu Lee (2021)
MACHINE LEARNING: As machine learning continues to advance and expand the boundaries of AI, its algorithms become better at emulating human emotions and intuition. Models are continuously learning and adapting, resulting in machines that eerily mimic human thought as well as creative and emotional responses. The gap between human and machine intelligence is shrinking and intelligent machines often exceed human characteristics – in speed, accuracy, durability, cost efficiency, and endurance.
Traditional machine learning can function with relatively small amounts of data, but the performance of the algorithms tends to plateau after a certain amount of data has been assimilated. The creation of neural networks brought about a fundamental breakthrough in machine learning. This enabled the approach called deep learning that involves training neural networks on vast amounts of data and building practical applications. Deep learning algorithms keep improving as more data is fed into the system. Deep learning produces a virtuous cycle of success leading to more success. More data leads to superior products, which brings more users, who produce even more data, which increases profits, which further expands the business model and allows companies to collect and exploit even more data. The gap between the winners and losers widens and become almost impossible to close. A few well-known examples illustrate this:
- Google is hard to beat in search engines. Microsoft tried competing by investing heavily in Bing but has failed.
- Facebook has cornered the market on social relationships and interpersonal interactions. No competitor has come close to overtaking it.
- Amazon dominates e-commerce. Even companies with huge financial backing have failed to supplant it.
DEVELOPMENT STAGES AND APPROACHES – A SCENARIO: Power Grids versus AI batteries: The giants aren’t just competing against one another in a race for the next deep learning. They’re also in a more immediate race against the small AI startups that want to use machine learning to revolutionize specific industries. It’s a contest between two approaches to distribute the “electricity” of AI across the economy: the “grid” approach of the seven giants versus the “battery” approach the of startups. How that race plays out will determine the nature of the AI business landscape – monopoly, oligopoly, or free-wheeling competition among hundreds of companies.
The “grid” approach is trying to commoditize AI. It aims to turn the power of machine learning into a standardized service that can be purchased by any company – or even be given away for free for academic or personal use – and accessed via cloud computing platforms. In this model, cloud platforms act as the grid, performing complex machines learning optimizations on whatever data problems users require. The companies behind these platforms – Google, Alibaba, and Amazon – act as the utility companies managing the grid and collecting the fees. AI startups are taking the opposite approach. Instead of waiting for this grid to take shape, startups are building highly specific “battery powered” AI products for each use. They build new products and train algorithms for specific tasks, including medical diagnosis, mortgage lending, and autonomous drones.
The Four Waves of AI: The complete AI revolution will take a little time and will ultimately wash over us in a series of four waves: Internet AI, business AI, perception AI and autonomous AI. The first two waves are already around us.
- Internet AI: The first wave began almost 15 years ago but finally went mainstream around 2012. Internet AI is largely about using AI algorithms as recommending engines: systems that learn our personal preferences and then serve up content hand-picked for us, and we become more addicted as it proceeds.
- Business AI: Business AI takes advantage of the fact that traditional companies have also been automatically labeling huge quantities of data for decades. For instance, insurance companies have been covering accidents and catching fraud and hospitals have been keeping records of diagnoses and survival rates. Business AI mines these databases for hidden correlations that often escape the naked eye and human brain. It draws on all the historic decisions and outcomes within an organization and uses labeled data to train an algorithm that can outperform even the most experienced human practitioners. This act of seeking out various correlations and making predictions is exactly what deep learning excels at. Given enough training data an AI-powered diagnostic tool could turn any medical professional into a super-diagnostician, a doctor with experience of tens of millions of cases, an uncanny ability to spot hidden correlations, and a perfect memory.
- Perception AI: Algorithms can now group the pixels from a photo or video into meaningful clusters and recognize objects in much the same way our brain does. Digitizing our physical world, learning to recognize our faces, understand our requests, and “see” the world around us. Third wave AI is all about extending and expanding this power throughout our lived environment, digitizing the world around us through the proliferation of sensors and smart devices.
- Autonomous AI: Autonomous AI will revolutionize so much of our daily lives. Early autonomous robotic applications will work only in highly structured environments where they can create immediate economic value. This means primarily factories, warehouses, and farms. But these machines are automated, they are not autonomous. While they can repeat an action, they can’t make decisions or improvise according to changing conditions. Giving machines the power of sight, the sense of touch, and the ability to optimize from data, we can dramatically expand the number of tasks they can tackle. Autonomous AI will come last but will have the deepest impact on our lives. As self-driving cars take to the streets, autonomous drones take to the skies, and intelligent robots take over factories, they will transform everything from highway driving to organic farming to fast food.
Differences in Approaching Global Markets: Silicon Valley giants want to directly introduce their products to global markets. They’ll make limited efforts to localization but will largely stick to the traditional playbook. They will build one global product and push it on billions of different users around the globe. Chinese companies are instead steering clear of direct competition and investing in the scrappy local startups that Silicon Valley looks to wipe out. It’s an approach rooted in the country’s own native experience. So, instead of seeking to both squash those startups and outcompete Silicon Valley, they’re throwing their lot in with the local businesses.
Source: AI Superpowers by Kai-Fu Lee (2018).
EVOLUTION OF AI DEVICES: In the decade spanning 2005 through 2015, the growth of smartphones positioned them for large-scale data collection. These devices, equipped with ever more advanced sensors, wireless networking, imaging, and voice capabilities, have been capturing much larger quantities of data than desktop or laptop counterparts. Social media companies grabbed this opportunity and became dominant platforms supplying a variety of services. Then the wearables space took off in a big way with several fitness-tracking watches quickly emerging with their functions eventually absorbed into smartwatches. The development of these devices established a new kind of data to mine at scale: health and fitness. In 2014, Amazon’s Alexa successfully overcame the first barrier to household penetration of smart home technologies. Amazon, Google, and Apple have made huge investments in home automation technologies and with appliance manufacturers to capture a plethora of data from households, including voice, video, temperature, soil humidity, air purity, Wi-Fi strength, how often someone opens the refrigerator, eating habits, and every other imaginable measurable by a gadget that’s connected to the AI hub at home. These voice interfaces will soon be embedded into ambient surfaces like ceiling and walls and furniture instead of being standalone devices.
Data privacy issues have not been adequately addressed vis-a-vis transparency on where and how all this data is stored and handled, who has access to it, and the relationship between the manufacturer and local law enforcement to govern its usage. From wearables, the industry is progressing to implants that physically fuse technology into our bodies and tinker with the creation of a new kind of augmented human being. These technologies may monitor our physiology and emotional responses in new ways and do so very accurately because of their proximity to the source of emotional responses in our bodies.
Each of the modalities [web browsers, smartphones, smart homes, wearables, implants] has broken new ground in creating a new kind of data about individuals that hitherto was unavailable or difficult to obtain. The more kinds of data that corporations can mine, the more complete the psychological and behavioral profiles of individuals can be built; hence the more sophisticated the ability of corporations to manipulate individuals and entire societies, by providing subtle nudges and cues through their every interaction with technology. Devices classified as wearables, implants, and smart homes are fundamentally changing not only how humans interact with machines, but ultimately influence the way humans think, feel, and behave.
Source: Artificial Intelligence and the Future of Power by Rativ Malhotra (2021)
CEDING CONTROL TO MACHINES: As AI systems advance and more of everyday life gets automated, the less control we actually have over the decisions being made about and for us. The current development track of AI prioritizes automation and efficiency which necessarily means we have less control and choice over the thousands of our everyday activities, even those that are seemingly insignificant. Our future living with AI begins with a loss of control over the little things. We’re not heading toward a single catastrophe but rather the steady erosion of the humanity we take for granted today. It’s time to see what happens as we transition from artificial narrow intelligence to artificial general intelligence – and what life will look like during the next 50 years as humanity cedes control to thinking machines.
Source: The Big Nine: How the Tech Titans & Their Thinking Machines Could Warp Humanity by Amy Webb (2019)
AI models and algorithms are touted as being objective and neutral, which is far from the truth. Biases of the people that control Ai systems enter at least two ways: through the data sets used for training and validation and via the precise definition of goals and objectives required for training machines. Machine learning models routinely make arbitrary decisions that the companies rationalize as an enforcement of community standards and community guidelines. But what are these standards and guidelines, who defines them, and how are they interpreted for enforcement? Social media companies have publicly established generic criteria like authenticity, misrepresentation, false information, respect for intellectual property, insensitivity to the dignity of others, hate speech, and so forth. These lofty principles are too broad and give the algorithms too much discretion. Though commendable at face value, in practice the implementation is inconsistent, biased, and in many cases, outright unfair. Worst of all, their algorithms lack transparency because the models that drive them are closely guarded secrets. Many posts have been rejected on spurious grounds that expose the digital platform’s cultural, racial, or ideological prejudice.
Source: Artificial Intelligence and the Future of Power by Rajiv Malhotra (2021)
USE OF DATA WITHOUT BREACH OF PRIVACY: Simply dissociating customers data from actual identities give insufficient protection and merely provides better optics in the eyes of regulators who do not know better. Using private data to make personalized shopping offers seems benign. However, the same advantage is afforded to a political group that seeks to influence votes, garner support for a cause, mobilize a revolt, or predict an individual’s behavior in a given negotiation or dispute. Such profiling of individuals is being used to develop psychological and emotional models likely to succeed, identify what threats and fears make the most effective deterrent, and enable other manipulative opportunities. These powerful impacts will pose serious challenges to social and political systems. But for now, AI is being hailed as a platform for utopia. The views of many futurists assume that society will proceed directly for the present social order to a superior one in the future. However, I foresee a prolonged period of disequilibrium in between, starting in the next decade and lasting until the end of this century. The youth of today will feel the brunt of this transition. By the time those born in 2020 are in their teens, the crisis of the present equilibrium will become obvious. And except for the lucky, sheltered elites that enjoy positions of power, most people will face uncertainty and turmoil.
Source: Artificial Intelligence and the Future of Power by Rajiv Malhotra (2021)
The unabbreviated version of the above can be found in the pdf document below.
| 8l_big_tech_understanding_ai_--_segment_8.pdf |