The Intelligent Automation : Learnings From a King's Life
Where and How NOT to Automate
Welcome! I hope this article finds you well. As usual, we are going to start with a disclaimer before writing this editorial.
We at Cyberscribble stand as both advocates and critics of technology and automation. While we cherish its potential for advancement, we remain vigilant in ensuring its application serves our collective betterment rather than leading to undesirable outcomes.
Having said that, today we are going to look into 'Intelligent Automation'. Let's begin!
King Croesus and Oracle at Delphi
Once upon a time, in the realm of Greek Mythology or more specifically in Asia Minor, there lived a King, who was so wealthy that he gave birth to the expression 'As rich as Croesus.'
Croesus was determined, ambitious, and had all the wealth in the world. Naturally, the next step for him in his ambitious legacy was to expand his empire. He eventually succeeded in doing so. At the height of his reign, he was the most powerful monarch in Asia Minor.
The ‘Fall’
Croesus had a good habit, he will always seek advice from Sages and Oracles before making any great decision. One day, he reached out to Oracle at Delphi and inquired about making war against the Persian Achaemenid Empire. Let’s imagine the conversation between them.
King: “Hey Oracle! if I wage war against Persia, what will happen? keep the response within 11 words.”
Oracle: “If you go to war, you will destroy a great empire.”
The Oracle, ambiguously predicted that if Croesus attacked Persia, he would destroy a great empire. Interpreting this as favorable, Croesus attacked, but the empire he destroyed was his own.
Sad!
Why did we tell this story ?
In modern days, we do have many 'Sages' and 'Oracles' living in our devices in the form of 'recommender systems' and 'artificial intelligence' such as 'Alexa and Google Assistant.' As of now, we use them to get advice on 'cooking,' 'traveling,' and 'entertainment,' among other things.
As AI advancement continues, we are seeing ‘Intelligent Automation’ incorporated vastly into our daily and institutional lives to help us make decisions . Some worth mentions are
Trading.
Trading intelligence systems which recommend and suggest our next trades.
Automotive.
robots on production line and prediction systems on supply chain shortages.
Healthcare.
Analysis , diagnosis and treatment.
Social welfare.
further bellow in this article.
Oh well, We can hear your mind voice going ballistic: 'We already know this! What's the catch?
The ‘Catch’
The problem with 'Intelligent Automation' systems that augment our decision-making these days is that the 'data' they are trained on and the 'algorithm' they use intrinsically have several flaws, which make them ill-suited as it is for certain industries and applications.
Flaw in ‘Data’
Bias : Data used to train these systems can contain inherent biases that reflect historical or societal inequalities, which the automation system then perpetuates
Incompleteness: Missing data can lead to incorrect or incomplete analyses, affecting the system's decisions.
Dynamic Environments: Data that does not reflect changes over time can become outdated, making the system less effective
Flaw in ‘Algorithm’
Ethical Concerns: Algorithms can make decisions that, while efficient, might not align with ethical or legal standards of fairness and equity.
For example, the classification systems, which are commonly used almost everywhere these days, usually work by 'separating,' 'ranking,' and 'categorizing' data of concern to make decisions, philosophically this is also a modus operandi of 'systems of discrimination'
Study : Intelligent Automation in Digital Welfare
In modern times, we have seen several nations rolling out automated intelligence in social protection systems. These systems are often presented as 'neutral' systems without human judgment bias, which can provide great coverage, reduce costs, enhance security, and, most importantly, detect frauds.
However, due to reasons such as "flaw in data" and "flaw in algorithm," the systems themselves are biased and discriminative.
Occurrence 1
Amnesty International discovered that an algorithm used in the Netherlands for detecting fraud in childcare benefits unfairly targeted non-Dutch nationals, assigning them higher risk scores. This led to increased investigations, benefit suspensions, and severe financial hardships for affected families.. The discriminatory system was eventually discontinued by the Dutch government [4]
Occurrence 2
Researchers in India found that individuals were erroneously declared deceased by a faulty algorithmic system used to establish eligibility for an allowance for those aged 60 and above.. Being erroneously declared deceased rendered people ineligible for support, and these individuals then faced an ordeal to prove that they were, in fact, still alive to be able to reinstate their allowance [4].
Lessons learnt
People who were employed to use these 'Intelligent Automation' systems inherently trusted the system's decisions so much that they ignored traditional sources of information, including their own judgment [4].
People who observed discrepancies and raised concerns, found that their concerns were ignored by their superiors due to a high level of trust in technology.
Decision Making: The Importance Of Diversity and Alternate Opinion
When making decisions , It is important to involve diverse sources, tools and multiple individuals to make a well-balanced decision.
Sources of Decision 21st Century
Experience
Education
Experts & Authorities
Research & Studies
Stakeholders
360 Judgment
Intelligent Automation Recommendations ( NEW )
Currently, many view ‘Intelligent Automation’ primarily as a means to cut costs and enhance process efficiency. While this perspective holds true in some scenarios, it often overlooks the necessity for additional training and the creation of new roles essential for effective implementation.
Moreover, ‘Intelligent Automation’ should be seen not just as a tool for efficiency but as an integral part of the decision-making process. This integration requires a robust pipeline with more, not fewer, skilled individuals to ensure balanced and informed decisions.
Without a comprehensive approach that includes adequate human oversight, organizations and nations risk facing the same pitfalls that led to the downfall of historical figures like Croesus.
Recommendations for Intelligent Automation
Implement an ‘Intelligent Automation’ compliance program which is supported by necessary standards
Establish a dedicated team to rigorously evaluate ‘Intelligent Automation’ rollout and champion diverse viewpoints within an organization.
An example Intelligent Automation Oversight Team structure,
Ethics and Compliance Officer for Intelligent Automation
Automation Auditors
Ethics Officers
Automation Risk Managers
Intelligent Automation Advisors ( Pro and Against)
For public institutions, establish independent and public oversight body over the use of automated and semi-automated decision-making systems.
Enhance, strengthen ‘Intelligent Automation’ decision making pipeline by adding more individuals as part of the decision making ecosystem.
An Example Intelligent Automation Decision Validation Team Structure:
Lead with a team of at least 3 members
Decision Engagement Specialist
For institutions and organizations which employ extensive ‘Intelligent Automation’ , it is highly recommended to form and operate a Decision Audit Office which encompasses
Chief Decision Auditor
Compliance Analyst
Ethical Violation Reviewer
Stakeholder Liaison Officer or Stakeholder Engagement Officer
As for us, it seems like we need more people to sustain the 'AI' and 'Intelligent Automation' era that is dawning upon us than fewer. Remember,
the next successful nation or organization in the world is not the one which creates a welfare state for people based on ‘Intelligent Automation’, but rather the one which employs them in various roles effectively to form the iron-clad ‘Intelligent Automation’ ecosystem.
Good to Keep the Human(S) in the loop!
Before we go, Did you all know ?
Philosopher Thales of Miletus worked as an Engineer in Croesus army during the military campaign against the Persians.
References
[1] Coresus