Benjamin Parry is…

Torn

Navigating everyday decisions by your guiding principles is hard.

Fireworks at Ryogoku Bridge, from One Hundred Famous Views of Edo, by Utagawa Hiroshige Source: artic.edu

Yes, we’re all somewhat hypocritical. And…

When this text appeared on one of Hidde de Vries’ slides at FFConf this year I immediately took a photo. In his talk, The future of the web is green, he poked at the environmental impact of the tech sector and our duty to make changes within. This has been something that has been nagging at me for a while.

Navigating everyday decisions by your guiding principles is hard. A seemly positive choice is often interwoven with a negative outcome. The reverse is also true: knowingly making a negative choice with the intention of offsetting the impact in the future. These scenarios and choices are a symptom of the complex systems we exist within.

Here’s a recent example. Two weeks ago on Thursday, November the 5th, bonfire night. A celebration of identity, history and tradition, in an exhibition of thunderous aerial explosions and colossal fires. This event defines the UK, in particular Lewes, the town where I live.

Lewes has seven bonfire societies. Each one lives and breaths bonfire each day of the year. It's bigger than football.

They bind the community together, giving people purpose. Two things that we'll come to rely upon in the coming years of inevitable climate breakdown.

This November 5th felt different though. The searing heatwaves and wildfires burnt into memory and reports of the impact on global air quality. The recent attempts to lower the decibels of fireworks and their impact on wildlife, ironically echoed by the new narrative for public transport "Share the space not the noise". With every fizz, crackle and bang I found my thoughts being torn.

This is a similar tension I have with AI.

I’ll be the first to admit, for some time, I’ve been loitering in the first few stages of AI grief. More recently though, I’ve started using it in my work and personal problem solving. Sometimes out of necessity and other times our of curiosity. Until more recently I’ve been steadfast in a blanket-resistance against AI in all of its forms. But this opinion hasn’t been based on balanced first-hand experience and perpetuated by confirmation bias.

Now that I’ve used it to solve a few problems, some far beyond my skillset, I feel like I can start to give an honest, balanced and informed opinion. What I’ve found interesting while using it, is the awareness of my own fallible shortcomings. Something I don't think I would have been open to or aware of were it not for my initial critical viewpoint. In that respect, my cautious optimism hasn't changed. I think the application has value but not a solution for every problem… hammer, nail, etcetera.

I am however, still very concerned of the environmental impact of large-scale data centers will have on eroding previous environmental efforts and catalyse further irreversible climate collapse.

My worries go beyond the environment. It has far-reaching impact into every facet of today's world and continues to shows up on a daily basis.

Through the lens of my career, every research product demo leaves me wondering "What's my tech exit strategy?" because it will be coming for my job, directly or indirectly, I’m certain of it.

The recent cyber security attacks on food and blood refrigerators, whether or not AI played a role, this technology provides a cheap and easy route to those who have the intention.

In this respect, it’s the people who push the buttons that scare me the most, not the technology itself. It’s the people who drive its future, influence its regulation, and profit from it; the Alman’s, Zuckerberg's and Bezos' of this world. When it comes to modern-day white guys playing with explosive fire, we’re spoiled for choice.

In the same way I don't expect Lewes bonfire societies to down torches and incendiaries for fairy lights and drums, I think we can approach AI usage with some measured intentionality.

We might lend from the software engineer principal of performance budgets, where instead of page load speed, file size and the end user experience, we’re focused on the energy and water consumption tied to our prompting. Credits and tokens are the closest thing we have to this currently but it’s otherwise obfuscated. In lieu of a transparent view of energy and water usage I’m going to double down on the pre-thinking and pre-work before turning to a prompt window.

We might also want to develop a set of clear principles for ourselves, the teams and organisations in which we work. A shared agreement of AI usage ‘this, not this’ statements that can help us hold ourselves and each other to account.

I'm trying to create healthy friction in my AI usage that goes toward lowering its impact on the world and hopefully lower the tension I'm feeling while using it. It's not perfect but it's a start.

Tags