Six years ago, covid-19 forced companies, schools, universities, and public agencies to adopt remote work — with no process of planning or discussion. In just a few weeks, video conferencing became the most widely used piece of technology, invading the daily life of millions of people around the world. That solution allowed many essential activities to continue, but it also produced exhaustion, an uncontrolled rise in workload — particularly for women, who added round-the-clock domestic and care work on top of their jobs — and a substantial increase in mental health problems across all ages. Do these negative consequences mean video conferencing is bad? The answer is a resounding no. What that experience teaches us is that rolling out technology without planning, without training, and above all without agreements within organizations carries enormous, entirely avoidable risks. Today's rising technology is AI: faced with its rapid arrival, we might hope history wouldn't repeat itself. And yet, as has happened so many times before, it isn't.

In recent weeks we've seen cases go around the world that are a clear symptom of a technology arriving and being implemented without explicit criteria or foresight: live monitoring of workers' computer use to train the AI of the company they work for; strikes demanding that workers share in the profits generated by this technology; and even the "distillation" of workers as part of the layoff process at Chinese companies. Each of these cases represents conflict, a loss of trust in the workplace, and frustration that affects not only the quality of life of the people involved, but also the productivity and the standards these same organizations hold themselves to. During covid there was no choice — we had to implement without planning and pay the consequences. Today there's no pandemic bearing down on us, but the breakneck pace of AI development threatens to squeeze out the space for reflection and review.

What internal agreements and criteria exist within a company or organization about the use of artificial intelligence in day-to-day work? How much room is there to even have that discussion? Perhaps the place to start is by checking whether anyone is asking these questions at all.

At Crítica we keep up a constant review of how we integrate AI into our own work. We're a diverse team, and so is the way each of us uses the different technological tools available in our professional practice. We know our priority is not to let down our own team, nor our clients: when we take on a project, we bring our humanity and ethical commitment to every task, proposal, and analysis. We work with an experimental approach to our processes, and that demands the full singularity, creativity, critical perspective, experience, and curiosity of every person involved. It's through this internal process that we've come to understand the value of AI-use policies inside any organization — how they connect to institutional identity and mission, what political and social positioning they imply, what commitment to work processes they represent, and what definition of excellence they express. We also look, from our integral approach to technology, at the technical capabilities and possibilities for carrying out that integration: what values and principles guide digital infrastructure decisions, and what the policy is on privacy and data use.

It all starts with asking the necessary questions. Our own experimentation lets us accompany other organizations as they chart their own path for integrating AI into their spheres of influence and their daily work. An artificial intelligence use policy can no longer be an afterthought. Today, the terms on which a company integrates artificial intelligence into its processes, services, or products should be part of its calling card — because it says a great deal about its priorities, its ways of working, and its identity.

We still have time to think, debate, plan, and move forward to minimize the risks and make the most of the benefits of implementing artificial intelligence. Today we can avoid tripping over the same stone twice. And that, fortunately, AI will not do for us — it requires, above all else, collective thought and human creativity.