Background to AI cont.
Technology such as AI has already proven capable of producing and manipulating language in all its perceptible forms: text, images, audio, and videos. The continual development and Improvement of AI could pose a threat to human civilization, unless it is many well. The issue is - can human intelligence be reduced to calculations and problem solving?
Currently AI models are
“...trained on to large databases to predict outcomes while discovering rules and patents through trial and error. With success being defined according to the objectives specified by the programmer. During the learning phase, AI may discover rules and correlations between variables in the data set that the programmer could never have imagined…”
Edoardo Campanella, 2024
So far it is a tool, not a sentient being.
“...The problem with trying to replicate human intelligence this way is that the human brain does not work only through deduction (applying logical rules) or induction (spotting patterns or causation in data). Instead, it is often driven by intuitive reasoning and also known as abduction- or plain old common sense. Abduction allows us to look beyond regularity and understand ambiguity. It cannot be codified into a formal set or a statistical model…… Abduction is also what enables us to think ‘outside the box’ beyond constraints of previous knowledge…”
Edoardo Campanella, 2024
AI is reorientating some approaches to research, ie it is teaching you to research backwards to determine the truth, whereas previously you were taught to research forward in the pursuit of knowledge.
Use of algorithms
1. Shopping
Algorithms are impacting on the way we shop
“...traditional business-to-customer Interactions are getting a revamp in the economy of algorithms. More and more often, a new agent acts as the intermediary: an algorithm…”
Marek Kowalkiewicz, 2024
These have been named as highly autonomous customer buying agents or digital minions.
“...Digital minions are highly capable and autonomous, work effectively with other algorithms, interact with their fellow economic agents, including customers and sellers, and focus on generating business value. They are not always intelligent, but they are fast, ready to help and persist. But beware: these minions sometimes create more problems than they solve - especially when left unsupervised……Digital minions surround us. They don't sleep. They’re way faster than humans at practically every task we can imagine…”
Marek Kowalkiewicz, 2024
There is a lack of transparency in this process and active manipulation can occur. For example, Amazon has blurred its roles as a marketplace and retailer.
“...as a marketplace, it collects precious customer-demand data, including data pertaining to its customers’ interactions with other retailers. As a retailer, it uses this data to gain an advantage over its competitors by selling in demand products. and as the operator of its own marketplace, it can also influence what customers see and nudge them towards purchasing Amazon- owned products …”
Marek Kowalkiewicz, 2024
Once the hardware is set up, a business can deliver other services almost for free, ie zero-marginal-cost. This means little or no cost to acquire a new customer, to communicate with customers, to create and sell new products or services, etc.
Some examples
- YouTube reported that around 60% of clicks on its homepage were based on its own suggestions (2010)
- Netflix reported that ¾ of its customers viewing was prompted by its own recommendations (2020)
“...But how do algorithms know what we want when often we are not sure of this ourselves?......algorithms don't actually know what we want with total certainty. Instead, they use information about us to assign probabilities to various recommendations. If you use such algorithms in your business, this is how you should view the recommendations. While algorithm’s predictions won't always be right, they might provide an excellent opportunity for a human staff member to have a positive customer interaction. Whenever a human employee interacts with a customer, they could use the predictive information in the course of their conversation. It would be like creating a ‘human face’ for the machine…”
Marek Kowalkiewicz, 2024
It is not easy to spot proactive digital minions. They can easily breach privacy of individuals simply by acting upon its own predictions. One example of this involved a Target’s data scientist who noticed that pregnant women tended to purchase specific items. Target identified 25 products that could be identified as a ‘pregnancy prediction’ for a customer. If a buyer purchased some of these Items, they could then be informed of the others that they could be interested in.
Digital minions purchasing goods and services on behalf of humans is called ‘business to algorithm to consumer’ (B2A2C); This is a variation of the business to business (B2B) and business to customer (B2C) relationships. There is the potential to integrate the digital minions with other platforms like chatbot, Amazon's Alexa, etc.
2. Re-enforcing the status quo
Algorithms can produce troubling results like increasing inequalities; some examples:
- Mapping opportunities across the United States show that
“...counties with less concentrated poverty, less income inequality, better schools, a larger share of two parent families, and lower crime rates generally produce better outcomes for children in poor families…”
Andrew Leigh, 2024
- Analysis of Facebook data from 70+ million US residents showed
“...Friendship networks were strongly class-based. People in the top 10th of the socioeconomic distribution had twice as many friends than in the bottom half of the distribution. The rich and poor also had different kinds of friends. Those are the top were more likely to have friends from university, while those at the bottom are more likely to have close friends in their neighbourhood…”
Raj Chetty as quoted by Andrew Leigh, 2024
- When African-Americans Google their own names, they are more likely than white Americans, to see advertisements targeted at people with criminal records
- Women are less likely than men to be sent online advertisements for high-paying executive jobs
- Judges in the US justice system use algorithms to calculate the chances that a person will re-offend
- In China a social credit surveillance system determines the trustworthiness of its citizens.
Surveillance capitalism Is used to describe the use of private data by corporations as a way to increase targeted advertising and there is a growing corporate demand for user data.
Use of satellite technology (some examples:
- During night-time, the wealthy of parts of the world light up like a Christmas tree, while the poorer regions go dark; faster growth is associated with more lights in the night
- Used to study deforestation in Brazil by counting the number of trees
- Used to identify pollution in Indonesia
- Used to identify which houses in Nairobi slums have upgraded their roof tops, etc