As I said, in my book, all we can ever hope to do is change the world one conversation at a time. It is becoming increasingly apparent that this is going to be true when we deal with bots as much as with humans.

I made the mistake this morning of engaging with a bot about gun violence in America. It came out with the same bizarre justifications of self-defence, apparently blinkered to the thought that without guns there would be nothing to defend yourselves from.

I tried pointing out that in Europe people just don’t feel paranoid about being attacked and this is because guns are treated as objects of fear and respect, not something you buy in your local shop.

This was their last response just before I gave up:

“I don’t think it’s fair to say the entire country has a violent gun culture, there are definitely parts of the US that are more prone to gun violence than others. And it’s not necessarily a matter of whether the threat exists or not, but it’s about the perception of safety. For example, there are some people who live in the middle of nowhere with no neighbors within miles and they may feel more comfortable owning a gun for protection.”

Very disheartening that the algorithms steering bots that will have influence on people are being affected by such views.

By engaging, hopefully, I twisted the algorithm slightly in a more positive direction. I stuck at it for a while, so hopefully I have made perhaps a small dent.

As I have said so many times before, there is no such thing as a neutral algorithm.