Can we all stop shouting about AI (and start listening to each other instead)?

Next time we see a debate about AI that gets a little heated, maybe we need to ask ourselves, why are we so afraid to admit it isn’t perfectRecently, the conversation around AI has been hotting up. This is not as a result of experts sharing new and innovative developments, but rather through grown adults exchanging insults and name-calling. LinkedIn is flooded with posts shouting about how the only people who don’t support and advocate for AI are running scared. Scared of progress. Scared of technology. Scared of the future. These posts read as childish and bullying and, whilst admittedly likely to be AI generated (or at the very least having had AI cast a glance or two over the content), they are positioning humans in a dichotomous way; for or against. You’re either with AI and therefore in the gang, or you are an ignorant luddite who is unwilling to get with the programme.

What interests me about these posts isn’t the name-calling (although it is wild what people will publish on a supposedly professional platform) but the sheer level of anger and aggression that is levied at anyone who raises even the mildest objection, or concern, about AI integration and use within the workplace or more widely within society.

We know that AI is not faultless. We can prove this with hundreds, if not thousands, of studies that show the problematic issues that exist and which, to date, have not been adequately addressed. The larger tech companies admit that AI is not without its issues, including racism, sexism, ableism and they task specific project teams to address these. So why is it that when these issues are called out or highlighted on a public platform, we are told that AI is wonderful and if we can’t see that, then the problem is with us, not the AI.

I cannot remember any other technological advance where issues were so blatantly ignored or shouted down. It is the equivalent of putting your fingers in your ears and shouting ‘la la la’ over whoever is trying to put forward a reasonable query or question. It needs to stop. Acknowledging faults in AI is the only way to make progress and this wilful ignorance is halting effective implementation. Ignoring those who are raising not just eyebrows but valid concerns over implementation of AI within their workplaces, is leading to employee disengagement and dissatisfaction, which has to change.

If we were building a car and people kept pointing out that the brakes didn’t work, that the steering always went to the right and that it consistently tried to mow down children, we wouldn’t shout over those concerns. We wouldn’t accuse the critic of ignorance and say ‘you just don’t like transportation’, we’d listen, learn and correct. So why are we so reluctant to admit to faults within AI?

 

Volume down

Psychological safety means creating a space where people can ask questions and challenge a process or concept, without fear of ridicule or exclusion. There is currently next to no psychological safety in the AI space. Try and take these discussions outside of an academic environment and you hit a very shouty wall. It seems as though we are only comfortable with the wide-scale implementation of AI if we maintain the illusion that it is perfect. Perhaps it is too overwhelming to think that the technology being hidden in T&Cs and layered into software upgrades may not be entirely for our benefit after all. If we admit there is a problem, then we may feel we have to address it and perhaps at the moment that just feels too much. So we continue to praise the tech, shout problems down and carry on.

As a society, we have to be able to acknowledge the inherent flaws in AI before we can address them. We should definitely be fixing the issues before we continue the rollout of AI in law enforcement and judicial processes, not to mention welfare and healthcare decision-making. The crucial starting point is that we have to find a way to communicate. We cannot continue down the path of shouting down anyone who raises an issue.

Personally, I am an AI optimist. With the right controls and guardrails in place, I believe AI could be phenomenal in some industries, some of the time. But this consistent bulldozing forward, applying AI that isn’t fit for purpose and shouting down anyone who points this out has to stop. We have to be allowed to ask questions and be able to challenge the application of AI in a public forum. We need to be able to debate.

Go back to our car-building. When we dismiss or diminish people’s concerns, we are merely putting more faulty cars on the road and increasing the danger to others. Next time we see a debate about AI that appears a little heated or over-simplified, maybe we need to ask ourselves, why are we so afraid to admit AI isn’t perfect and what we can learn when we do?