Big Data, Artificial Intelligence and Freedom - Cyber Dystopia and Beyond

Index of Blogs and Courses

'Artificial Intelligence is more profound than fire or electricity' - Sundar Pichai (CEO Alphabet)

'Country leading in AI in 2030 will rule the planet until at least 2100' - Brookings Institution Report

'If you have enough data about me, enough computing power and biological knowledge, you can hack my body, my brain, my life, and you can understand me better than I understand myself' - Yuval Harari (Historian)

Introduction:  Recently, there has been much disquiet expressed about the uncertain world that exponential rise in technology might create for our civilisation. Previous technological revolutions have, on the whole, helped us humans in creating healthier, more comfortable living conditions.  The trio of technologies; bio-, nano- and digital-, mostly developed in the second half of the 20th century, have transformed life and augmented what we can do in a big way - generally for the better.  The two most damaging aspects of technological development in the 20th century have been human population explosion and ecological collapse.  To control these global problems needed co-ordinated global action and co-operation among the nations of the world.  This was the missing element, and humans have let the situation get out of control with dangers of many unpleasant consequences in the future. 
In the 21st century, digital technology has the potential of affecting our lives in ways that was not possible in the past.  Our privacy, freedom and human rights are all at risk.  There is a fair possibility that even if nations of the world co-operate in tackling such risks (unlikely from past experience), it may not be possible to contain serious damage to our human rights. 

What is AI? - Artificial Intelligence (AI) or  intelligent machines is the concept that digital devices (computers) can make decisions based on the data being fed and are essentially able to 'think'. During the 20th century,  based on a set of instructions (algorithm) computers could perform certain tasks more efficiently - this is artificial narrow intelligence (ANI).  A chess playing computer that beats the best players is an example of ANI in action.  ANI capabilities have been increasing exponentially in terms of number of computations a computer can do per second.  The question is when will the computer match the capabilities of the human brain?  Ray Kurzweil thinks it is 2029 - others seriously doubt that, but most experts agree that it will happen in not too distant future - by about 2050. 
It is convenient to think of three stages in the development of AI - the slide explains this:

The graph explains the situation further: 

For a transition from ANI to AGI, machine learning (ML) is an essential first step. ML enables a computer system to automatically learn from real-life data and improve from experience without being explicitly programmed - the computer learns automatically without human intervention or assistance.  The amount of input data is important - the more the better - and current trends in collecting large quantities of data (Big Data - see below) is helping AI systems to progress rapidly.
In this blog, I shall restrict the discussion to the near future spanning the next few decades and try to figure out what might happen as AI evolves towards becoming human level intelligence (AGI).  

However, it is good to note that there is much ongoing discussion about the benefits of AI, how to control AI, dangers posed by ASI to human race etc. ASI will have far superior intellect than humans and will not be bound by rules made by humans. It is a moot point if ASI will act to safeguard human rights and support 'prone to irrationality' humans.  These are uncharted territories. 

Threats to Our Privacy and Freedom - Digital technology has been very useful in making our lives more interesting and industries more productive.  Over the past 40 years, technologies based on ANI and some early modest capabilities in AGI have transformed our lives beyond recognition.  This is reflected in the way the valuations of tech companies have rocketed - they now occupy seven out of ten top positions with many at trillion US dollar valuations. The slide also shows how rapidly the valuations of the tech companies have grown, and continue to do so. 

Of the seven leading tech companies, five are based in USA and two in China. The size of a company determines the power it can wield - they can seriously influence regulatory environment and use their vast funds to undermine competition and new start-ups. Tech companies also have global reach.

In order to understand why the rise of tech companies might on the whole be negative for our privacy and freedom, we first need to understand what 'big data' is.

Big Data:  
Large amounts of data is generated and collected every second.  To understand the magnitude of this, consider that you have 1 GB (109 bytes) of data available for use on your smart phone every month.  The global data is estimated at 33,000,000,000,000 GB or 33 Zetabytes (33 x 1021 bytes), is increasing rapidly and is expected to reach 175 ZB by 2025. The Volume, Variety and Velocity of the data underpins the development of artificial intelligence. 
Data about us is collected in many ways - smart phones, social media, smart home appliances, consumer products etc.  Much data about us is combined and shared among commercial, government and other institutions without our consent - in fact lot of the time we are not even aware that this is happening. 
Recent advances in face- and voice- recognition software allow you to be recognised in real time.  Such AI is already used by civic authorities in many countries to monitor peoples' movements - generally without any legal authority.   
These extremely large datasets (Big Data) enable computers to reveal useful hidden patterns, market trends, customer choices etc. leading to more informed decisions.  Big data may also be analysed to extract information about multiple aspects of human behaviour on an individual and collective levels. 
Tech companies 'own' and control Big Data.  With increasing capabilities of AI, they are in a strong position to exploit the information.  Let us look at it in more detail.
AI and Threat to Privacy:  One can consider privacy in two ways:  
1) Physical privacy concerns with your location, information about friends and family, biological (genetic DNA) information - medical records, financial information etc. Such information is what Big Data collects and stores. For example, Google has just sent us a log of all the places we had visited in the past 12 months.  It also keeps record of what we have been searching on the internet (our shopping habits).  Details of our bank accounts, medical history etc. are all digitised and are available to anybody who has the authority to access these.  
2)  Privacy of Thought: was indeed so during the 20th century.  With the ever increasing speed and scale of AI, computers can analyse Big Data to profile our thoughts with increasing accuracy.  For example, the way you drive your car may be analysed to deduce your emotional states such as anxiety, alertness, distractions etc. Similar information may be obtained from your keyboard typing patterns or your speech, the way you walk etc.
Even more worrying (and sinister) developments of AI have been the way data about you and those around you may be combined to deduce practically any aspects of your thinking like your political views, sexual orientations etc and can have massive influence on your credit worthiness, employment prospects and much more.  
Output from AI is not 100% accurate all the time - it looks for the most likely result.  AI does make mistakes and this might be catastrophic for those who are wrongly targeted for actions and thoughts that they are not guilty of.  There is no appeal allowed - AI results are final.  Authorities now use AI routinely for facial recognition and gait recognition to monitor communities suspected of challenging thoughts.  How many people are wrongly convicted or harassed is a good question.

BUT THINGS ARE MUCH WORSE: It is beyond doubt that within the next decade, AI will help authorities to effectively monitor and control their people.
A far more sinister possibility is the dominance of China and USA in their AI capabilities and the dominance of privately owned tech companies to effectively enslave the populations globally.  
I had analysed this aspect of AI in my September 2018 blog  and refer you to the second half of the blog for details. Enjoy.