Nothing artificial about intelligent use of print

It is a year since artificial intelligence broke through into the public sphere with the release of what was then the latest version of Open AI’s ChatGPT large language model. All sorts of predictions, dire or otherwise, were made about its impact and the dawn or death of civilisation. Those have not come to pass, though artificial voice mimicry and other aspects of the concept technology have been astounding. In print we have been promised applications that will render unexpected breakdowns obsolete. Instead we will get the unexpected visit from the engineer to fix the problem that the software has identified will occur later that day. Many long in the tooth digital printers will have experienced this without AI in the background.

That is where the beneficial impact of AI will be: solving issues before they become problems; handling transformations from one colour space to another; handling the profiling of a new paper and optimising the scheduling and flow of work through a plant. There will surely be new types of print job to follow, though with the exception of some ChatGPT created novels, little seems to have been produced to date. Which is a shame because in high speed production inkjet there is the tool to go hand in hand with AI content generation.

The report on the take up of inkjet print from the US points to incremental improvements from its use – better productivity, reduced waste and work captured from other print technologies. But no real breakthrough products that could not have been produced or conceived of before. They will come. In the meantime the arrival of AI that scrapes the internet to stir up new words and images is becoming the monster that feeds on itself. A recent Reddit discovered that the majority of images of baby peacocks were not of the bird, but of AI’s woefully inaccurate attempts to depict the chick. The product of AI represented digitally cannot be trusted it seems, while trust remains the key attribute of what is printed.