- The concept of “technical debt” has emerged to describe the gap between a programmer’s standalone skill and the features implemented by GenAI models. High technical debt exposes the user to expensive problems when GenAI comes short due to their inability to fill the gaps from lack of understanding or proficiency to apply understanding.
→ Therefore, GenAI should only be used in a capacity to help speed up or control undertakings the user could already take on on their own. A good example for responsible use would be debugging programs written by the user. GenAI will be more helpful towards pinpointing the breaking point, which can then be fixed thanks to the understanding of the program built whilst developing.

In short, GenAI should serve more as a steering assist than run a project on autopilot in my view.

• Some would argue that products entirely generated by GenAI tools from their prompts should also count as art they have created. While the contribution is certainly there, someone commissioning a piece of art is by no means the artist in a conventional exchange, therefore I maintain that AI Artists are more commissioner than artist.

• Having said this, the main purpose of “art” in all its forms, is the expression of a message with layers of subtleties and undercurrents intended by the creator for its consumers to interpret a certain way. It is the artist imprinting a snapshot of themselves onto their figurative canvas.
  A commissioned artwork may be intended to convey a certain meaning, however the interpretation of the expressed message is in the hands of the artist to configure. An honest artist will attempt to align their interpretation with that of the sponsor, however they could add personal imprints meant to serve their own agenda if they choose. This is a favourite method for artists in represive systems to exert criticism of sensitive figures for example, e.g. Jean de la Fontaine’s fables criticising the absolute French monarchy of the 17th century.

• By delegating the execution of a message to a generative model, prompters therefore lose the personal touch of the product, therefore undermining its quality as a work of art.

• If the intention of the message is an impersonal communication, GenAI can be a perfectly suitable medium. For anything aiming to express a personal message or opinion, AI-generated products risk losing their authenticity and intellectual depth. Therefore, GenAI is not suitable to express a personal, deep message.

• Due to generative models producing aggregates of information gathered from other sources, many academic institutions consider undisclosed usage of GenAI in works a form of plagiarism, or at least academic dishonesty, a view I subscribe to. For intellectual honesty, use of GenAI must be disclosed, and ideally accompanied by a description of how it was used in a project.

• Current models functioning off transformer architecture trained on existing data do not have an inherent creative component to it, merely the ability to combine its training data to a wider range of outputs.

• Models trained on Human-generated materials that have not been cleared for factual accuracy and intellectual honesty are prone to ingesting (deliberately or accidentally) misinforming data, which can in turn be included in the outputs. Furthermore, some information is subject to layers of context that can be misinterpreted (satire is famously difficult to discern algorithmically for example).
  → GenAI in its current form should never be considered an authoritative source of information or creativity.

• GenAI models and the Infrastructure erected to maintain and further develop them, as well as required adjacent systems, notably dedicated data centres, cause proven

  • ecological (water shortages, light and noise pollution, sometimes replacing existing ecosystems in irresponsible jurisdictions),

  • economic (job instability rising sharply since 2022, rising utility costs near data centres from expanding resource usage, unlawful use of copyrighted materials and contents explicitly designated against AI training for improvement of models without compensation), and

  • societal (enabling of deep fake contents putting reality into question, and making fraudulent practises increasingly accessible to bad actors, proven detrimental effects on cognitive abilities of children and long-time adult users alike)

  damage

• GenAI companies are, in some cases, led by individuals allegedly actively promoting

  • mass surveillance (Flock cameras facing intense pushback, an attempted integration into Ring [Amazon] cameras in Feb. 2026),

  • political extremism, associated hateful ideologies, and democratic backsliding (no names for legal reasons), as well as

  • increasing usage for automated, offensive (non-supporting) military activities.

  While not every tech executive fits this profile, there is a trend in the indirect direction society is being driven towards, by the people who benefit immensely from this sort of technology.

• While there are of course many advantages to GenAI and associated technology for larger society, and some of the criticised measures on paper serve a helpful purpose or have caveats to mention, e.g.

  • the increasing autonomy of cameras being definitely helpful in detecting and preventing serious criminal activity, and

  • the rising popularity of extremist politics being noticed before the popularisation of this technology,

  GenAI does at least exacerbate these issues, and at times actively or indirectly contributes to their proliferation.

  → For all the reasons outlined above, I believe it is an ethical imperative to reduce the use of such technology to the absolute, bare minimum, and abstaining from use entirely if at all possible.

  → On a more positive note, like every problematic technology in the past, I believe it is just a question of time and political will until GenAI becomes a net positive for society.

  • There are already projects undertaken by some of the best engineers on the planet, alongside the most vocal activists, to address the harmful effects of data centres,

  • Institutions will, in time, find ways to regulate applications of such technology with more robust enforcement mechanisms against unlawful usage,

  • Schools will adapt to teach their students how to think for themselves alongside such tools, and workplaces’ interplay between AI and Human employees will eventually stabilise again at some point.

  Just like nuclear technology has been harnessed now for its boundless energy, like factories have maintained and expanded on their incredible productivity from the industrial revolution to improve working conditions (not everywhere sadly), GenAI is just another revolution of the way we work and live, to which we will adapt one way or another.