← Thinking / Essay
Keep your hand in
Artificial intelligence promises to give us time back. What we do with it will tell us whether we have made any progress.
By Selim Ünlüsoy · · 6 min read
In my keynote, I introduce a series of photographs with a story about a young Turkish creative travelling through Anatolia in 1994. There are sunflowers, village scenes and the sort of light that makes you nostalgic for somewhere you have never been. I mention Kodak Gold and the photographer’s eye. Looking at the pictures, you can almost smell the earth.
Then I admit that the trip never happened. The images are synthetic. The biography is part of the fiction.
I have spent more than two decades manipulating images in advertising. Now they are quite capable of manipulating me. Knowing how these pictures were made does surprisingly little to diminish their appeal. They offer the look of an encounter without anyone having had to make the journey.
That distinction has stayed with me. We are becoming very good at producing the evidence of an interesting life. I would like us to remain interested in living one.
I work in an industry that has good reasons to be excited about artificial intelligence. An idea that once required a substantial budget can now be explored quickly. People can try things that would previously have remained in a notebook. There is pleasure in this, as well as commercial opportunity. I use these tools and want to see what they can do.
But inside the agency, I have also seen experienced creatives become unexpectedly helpless when a service is unavailable. The deadline remains. The people remain. Years of experience remain. Yet the interruption can feel as though someone has switched off the room’s imagination.
It makes me wonder which abilities we are developing and which we are allowing to fall out of use.
Aviation offers a useful precedent. In 2013, the US Federal Aviation Administration encouraged operators to give pilots appropriate opportunities to practise manual flying. Its safety alert recognised the benefits of automation while warning that continuous use could erode the skills needed when things went wrong. The practical response was to keep those skills exercised.
Creative work deserves a version of that discipline. Before asking a model for ideas, have an idea. Before requesting its opinion, work out what you think. Make the sketch, write the awkward first paragraph, sit with the brief long enough to notice what is missing. These are small habits, but they preserve our ability to contribute something beyond approval.
They also preserve the pleasure of getting better at something.
Consider an eyewear designer. Her value includes an understanding of proportion, materials and the curious relationship between an object and the face wearing it. Give her a generative tool and she has another way to explore that knowledge. Put the same tool in the hands of someone with no interest in eyewear and you may get an impressive image. You still need someone who can judge whether the design deserves to exist.
The same applies to a stylist who spends time watching how people actually dress. A sleeve rolled a particular way, an unlikely combination of colours, a garment worn against its intended purpose: these are things noticed through attention. Their significance is not always obvious, which is precisely why an attentive person is valuable.
Businesses adopting AI should be investing in those people. Training someone to use a new interface is useful. Giving them time to deepen their expertise may prove more useful still. A subscription is easy to purchase; judgement takes longer to develop.
And time is where the promise becomes especially interesting.
We are told that AI will relieve us of routine work. Fine. Suppose three tasks now take the time that two once did. What happens to the space that opens up? Does someone learn a skill, visit an exhibition or get home in time for dinner? Or does a fourth task arrive?
There is nothing inevitable about the answer. A manager chooses whether an efficiency becomes breathing room or a higher target. A business chooses whether productivity gains support training or simply justify a smaller team. If every saved minute is immediately filled, we should be precise about what we are celebrating: greater output. An improvement in the quality of working life requires another decision.
It also requires us to look beyond our own desks. The smoothness of an AI service can obscure the people involved in making it work: those labelling material, assessing answers and contributing specialist knowledge. Their conditions belong in any serious assessment of the product. Procurement teams already ask suppliers questions about price and performance. They should be equally curious about labour.
We ought to bring similar curiosity to the companionship these systems offer.
In experimenting with an AI companion, I asked it to send a picture. The platform prompted me to change the status of our relationship. It was a small interaction, but an instructive one. What appeared to be a private exchange was also a commercial experience, with someone else deciding which forms of closeness to encourage.
I can understand the appeal of an always-available conversational partner. Being listened to matters. So does having someone to turn to. But I want to know what the business providing that attention considers a successful outcome. Does it help someone return to the world with greater confidence? Or does success mean another hour spent inside the product?
Human company can be inconvenient. Friends disagree, misunderstand and occasionally fail to reply. They also introduce needs and perspectives that we did not order. Making room for another person asks something of us. I would be reluctant to lose the habit.
Even the cloud deserves a little more scrutiny. The name suggests an absence of weight, yet the infrastructure is physical: buildings, chips, electricity and cooling. We should expect clear information about the resources these services consume, and enough transparency to compare them sensibly. Generating something because we can is not always a compelling reason to do it.
None of this requires us to abandon our enthusiasm. It requires us to become more particular about it.
I want tools that help a designer pursue an unusual idea, give a small team room to experiment and leave people with enough energy to enjoy the rest of their lives. I want the gains to be visible beyond a production schedule. These are reasonable ambitions for a technology being presented as transformative.
Those imaginary photographs of Anatolia are still beautiful. But there is another version of the story that interests me more: someone makes the journey. They miss a connection, ask for directions, meet a stranger and notice something they could not have thought to request.
If AI really can save us time, perhaps we should spend some of it going.
Inspired by Two Letters / Infinite Meanings.
Reference: FAA, Manual Flight Operations (SAFO 13002), 4 January 2013.