EU flag

AI Act a first step toward recognising reality

New EU regulations on AI are to be welcomed, but we need to think more about the technology’s failings, says Jason Walsh
Image: Shutterstock via Dennis

15 March 2024

Passing into law on Wednesday 13 March, the new European Union Artificial Intelligence Act (EU AI Act) is a clear recognition of the pace of AI development, and its risks, but it is also worth thinking about AI’s increasingly obvious limitations.

The new rules covering general purpose AI models like ChatGPT will come into force in 12 months, with a two year delay before the rest of the Act is enforced. Emotion recognition in the workplace and schools, social scoring, and predictive policing based on profiling will all be banned, while other uses will be assessed according to a risk score.

The Act does not come as a surprise, having been in the pipeline for some time. However, its adoption comes at an interesting time: hype about the transformative potential of AI has settled down, with more reasonable expectations, such as its ability to summarise meetings, replacing fantastic visions of a fully automated future. 




Remember, AI’s boosters claim that the technology will spark a new era of human creativity, by freeing humanity from the hard slog of tedious and repetitive labour.

The problem is, we’ve heard this before.

In a 1981 television interview, the late Apple co-founder Steve Jobs made the following claim for computers: “One of these devices can free a person from many of the drudgeries of life and allow, really, humans to do what they do best, which is to work on a conceptual level, to work on a creative level.”

In a trivial sense Jobs was right: computers have aided with a great many tasks. However, many of us today working in offices, often functioning as adjuncts to spreadsheets, would be forgiven for asking if they are being creative or, indeed, if creativity is truly welcome at all.

Machines, by their very nature, are defined by their systems and processes, which is inherently limiting. While the raw calculating power of computers has been, and still is, genuinely transformative, these very systems and processes have a tendency to ossify and leave humans working to the tune of the systems rather than the other way around.

It is also interesting that those who shout loudest about the ‘creativity’ of AI systems, as opposed to their utility for technical tasks, seem overtaken by a feverish desire to rob humanity of creative tasks. It hasn’t exactly worked out, though. Interestingly, much of the visual art produced by AIs, even as its technical competence improves, is of no aesthetic or conceptual interest at all. In fact, it tends exceedingly strongly toward the repetitive and kitsch. It is, it seems, rather obviously process-bound.

Still, there can be no doubt that plenty of organisations are willing to pump out high volumes of low-grade work and if AI can assist them in unleashing the flood then they will use it. The risk is not just that our culture would be diminished, but our very lives.

The EU AI Act will do nothing to stop businesses and governments from bending themselves out of shape in order to suit computers. Its existence, though, is a weak but nonetheless important recognition that technology can, and should, be put at the service of humanity, not vice versa.

Read More:

Back to Top ↑