Since the dawn of humanity, humans have always sought to do things with less and less effort. This may seem frivolous, but it's true.
The effort made is found in the path taken to achieve an achievement, which is what prevents it from having to be reproduced later by the users or beneficiaries of the invention.
Artificial Intelligence (AI) is more of that, but in a more advanced version. science, where mathematical logic and computer science converge.
This post aims to emphasize what AI can bring to the table, and I prefer to be positive and interpret it as something that adds, not subtracts. Reflection should always be the first exercise we all undertake when faced with any event we face.
Alan Turing, a mathematician considered the father of AI, asked himself the following question: “The question is not whether machines think, but whether they do it better than humans.”.
Machine Learning and Deep Learning
He machine learning (machine learning) and the deep learning (deep learning) are two concepts that are linked to AI, and that surprised us before the latter.
He machine learning It involves machines (computers) learning by themselves through data analysis and the identification of behavior patterns, which allows them to predict behaviors without human intervention by developing generic algorithms. In other words, it's the science by which computers learn from data.
This is very familiar to us, as we are all targeted by machines when they track us, for example, on commercial websites we visit, and then offer us objects or products that may be of interest to us based on the information these machines have obtained about us and subsequently analyzed.
On the contrary, the deep learning It is something more "deep", as its translation suggests. And in this case, the machine It approaches the human brain in the way it processes the information it receives, and all of this through artificial neural networks, which achieves a more precise prediction with respect to that achieved in the machine learning.
In essence, what these sciences aim to achieve is to obtain the greatest amount of data or information to subsequently act upon it. But this means that the human being from whom the information or data originates is also affected, which on the one hand is positive but on the other can be dangerous, since having a vast amount of information can sometimes lead to control and power over something or someone. And this is the true danger these technologies pose.
Application of Deep Learning and Machine Learning to Compliance
If we focus on Compliance, and especially on ethical, honest and principled conduct beyond what is strictly related to Regulatory Compliance, we find that these sciences, especially deep learning, can be dangerous and cause considerable risks, especially if used without caution and prudence.
There is no doubt that Hobbes's famous phrase “man is a wolf to man"It gains more depth as humanity advances. And all advances, in addition to having their positive side of progress, innovation, and contribution to humanity, also carry their part"bitter"which translates into anxiety, uncertainty and danger for human beings.
And this is where what, from my point of view, should guide all software programming (algorithms) that will serve as the engine of machines or computers designed to process countless data comes in.the oil of our time> in order to improve people's lives, and not the other way around: I'm talking about ethics.
Ethics is the central axis of these algorithmic engines. Without it, neutral software cannot be conceived, one that doesn't tend to lean toward one side or the other to the detriment of a third party (person, company, society, etc.). Ethics must be involved throughout the entire process of information gathering, processing, and drawing conclusions, because only then will the decisions derived from the analysis of the data thus managed be, at the very least, ethical and respectful of people.
At the heart of the matter, we are talking about biases and partisan intentions. in the scientific chain, which would certainly lead to unfair actions. And injustice would be another negative consequence of a poor application of these sciences. of progress, since what they pursue is to obtain information to improve people's lives and not the opposite, which would be to cause harm.
We can imagine an AI-based risk identification tool that would be biased in such a way that it would partially consider risk assessment from a completely adulterated probability perspective, based on data chosen with the subjective criteria of the person who defined the algorithm, with a hidden intention that has nothing to do with transparency and best practices.
Of course, if this were to happen, we could always resort to the famous RBA or Risk Based Approach (risk-based approach) and act with caution and prudence before validating the results of that analysis.
For this reason, among others, we must show confidence in the existence of the Compliance function, among other missions, to prevent risks or situations that could cause harm to the Organization.
Consequently, it is important to be mindful, specifically in the area of Compliance, of all data processing and its reporting, providing conclusions and measures to be taken into account for controlling processes and, ultimately, business activity. Any bias or suspicious indication of discrimination in the process or outcome must be corrected immediately.
It is of unquestionable importance to reflect and analyze the results obtained and ensure that they do not present any anomalies or discrepancies with what we have assumed to be ethical and fair.
I would like to conclude with a quote attributed to Bill Gates, very much in line with what has been discussed in this post: “The first rule of any technology used in a business is that automation applied to an efficient operation will magnify the efficiency. The second is that automation applied to an inefficient operation will magnify the inefficiency.”.
If you want to know more current news about Data Protection Audit, Risk Management and Cyber Compliance, visit our blog.
