In April, I published the article “What type of AI usage?” In it, I observed that many companies could use a slightly more outcome-driven approach to AI to improve their impact. No tool is great in itself; it needs to solve a problem. This is utilitarianism – and it applies equally to AI, which is a tool.
The Industry strikes back Link to heading
After publishing my article, a good friend reached out. This friend warned me that criticizing AI could “negatively impact my career”.
I explained that my article simply took a balanced look at AI. AI has its uses, when used thoughtfully! Still, even remotely criticizing AI could do me a lot of harm, they argued.
My friend had recently been speaking to a lot of investors. Every investor asked, “Do you use AI?” Investors wanted “AI” regardless of the outcome. This means boards required “AI” irrespective of outcome. My friend argued my hireability would diminish if I had a blog post “criticising AI” somewhere online.
As product builders and engineers, we live in interesting times when we must not challenge the tools we are given. Or worse, we are not allowed to pick our tools ourselves. Where is the trust that we ourselves can pick the best tool for the job?
I am by no means criticising my friend here. They cared for me. They did not silence me, but I think our industry tried.
AI at all costs Link to heading
Another friend used to work as a Staff Engineer at a fairly large tech company in Stockholm. A smart friend I look up to. One day, the executive board announced that “all quarterly goals were to be dropped.” The company would instead “focus on AI” for the next six months. No specifics.
My friend asked the leadership, “What type of AI are we talking about here?” People were annoyed. What question was that? “All”, they answered. The whole company was in turmoil with random AI initiatives all over. Initiators of any “AI” project were rewarded1. Irrespective of impact.
My friend asked what the expected outcome was for specific AI initiatives, but got no answers. He was just considered annoying. People thought “AI” was the outcome.
Six months passed. Most initiatives were failures. A lot of wasted person-hours.
Reading between the lines, my friend was asked to stay silent. Not by a specific person, but by the company culture (as an organism).
Silenced cultures have less impact Link to heading
If we want to have impactful, successful companies, we must build a culture that rewards being challenged. I understand it has friction. Getting challenged can be tough. But it is for the better.
We live in dangerous times if we are not allowed to ask hard questions and challenge basic assumptions – be it about AI or something else. This applies to companies, our industry, or – God forbid – the world as a whole.