Skip links

All The Things Computers Can’t Do

Blog. We are forgetting the nature of humans, when we are applying the methods of computer problems. Human ability to think in the whole can never pass on to the computer whose language is binary encoding of data. And human language and the sociality of being together human-to-human cannot be reduced to binary encoding.

The usage of words as ”data”, “algorithm”,  and “artificial intelligence” has recently exploded and so has the promises told about these. The promising potentials seem enormous and areas still unexplored by the algorithmic-power are just waiting for the technology to solve all its problems – including problems in traditional human-human situations.

However, I will argue that we are forgetting the nature of human when we are applying the methods of computer problems. This is not an argument against the use of modern computer technologies, but I will try to explore the limits of the computer potential – by recognizing the human potential.

This text is inspired by Peter Kemp’s book “Det Uerstattelige – En Teknologi-Etik”.

The Potential of Computers
Originally, computers were persons with the working description “one who calculates” and were employed within the area of mathematics. Therefore, the first electronic computers were programmable electronic machine that performed high-speed mathematical or logical operations. These machines were still applied to the fields of mathematics and nearby, such as engineering and accounting.

The electronic computer is, per definition, a machine for processing electronic data in the form of formalized symbols in a binary encoding. The binary encoding is the language of the computer: It takes input in a binary encoding, doing binary logical operations and manipulations, and outputs as well something in a binary encoding. The rest is done by humans designing the algorithms and assigning meaning to the input and output.

The computer is processing data. Data to which we are assigning meaning and then calling information. Contrary to “traditional” technology, computer technology does not transform energy and raw material, but information which we use for communication. The understanding of modern technology as primarily a tool for the human’s relation to nature – or what Karl Marx called a metabolic interaction – does not count for information technology, which first and foremost is a tool for human’s relations to one another. Hence the information technology becomes a social technology meaning technologies which are organizing and distributing communication and work among humans.

If we are trying to develop social technologies as if humans were merely objects that can be worked and manipulated, used as input and outputs without problems, this will according to Kemp be perceived as oppression.

How come? Because we are trying to replace the human reason by computational power, by reducing our language to computer language – to algorithmic code.

The Human Reason
Peter Kemp is, clearly inspired by the philosopher Immanuel Kant, describing the human reason as the unification of rationality and imagination, as it is the ability to use the reason in the service of imagination. “In other words, the ability to put means and goals in context – to think in the whole.” (Peter Kemp pp. 230). And this ability to think in the whole can never pass on to the computer whose language is binary encoding of data.

An ethical problem arises, when we think that humans can unfold its skills and abilities with the rationality alone without unfolding imagination. Our language and the sociality of being together human-to-human cannot be reduced – or even translated – to binary encoding. This requires an extreme simplification, and this is what is happening when we are unfolding skills and abilities with the help of a computer. A completely automated society will result in radical changes of social communication systems towards an extreme depersonalisation and a risk of total monitoring and control.

This leads us a big ethical question: What is the principal limit for the electronic automation of our societies and lives, which will be ethically acceptable? Where is the limit for our treatment of each other with algorithmic methods?

By using computers in situations which traditionally has been human-human interaction we will miss out on a lot of informal information and being together. By only using the rationality, and not imagination and the ability to “think in the whole”,we are incapable of putting ourselves in the place of others. A computer cannot “think in the whole”. A computer cannot put itself in the place of a person. A computer cannot be sensitive and understanding. It can only “understand” binary encoded input.

The nature of the problems
We have witnessed how the computers have solved problems with great accuracy in fields of applied mathematics and automation of tedious administrative tasks. I believe that it is especially the accuracy and (what appears as) correctness that seems appealing to our desire to expand the usage of the computer to non-mathematical areas. Imagine a computer as a social worker, who always is making the “right” decision, and furthermore at a speed, where human colleagues cannot keep up. But putting the computer in the place of the human social worker, we will lose the human touch on the human problem.

We need to remember that the human’s problems are not solvable by the same methods as computer-problems. The problems are of a different nature!

And thus, life with the computer holds a danger that the personal life dissolves. Of course, this does not mean that the computer must be abolished. But it means that everyone who has to do with computerisation of society has a responsibility to make this happen as humanly as possible, by only applying the computer methods to computer problems.


Ida Marie Schytt Lassen is a philosophy – and computer science student at Aarhus University