I’ve largely avoided AI chatbots for the most part of the AI boom as I’ve seen no use and the majority are just not needed, I mean Snapchat, really? These bots are very prone to going a bit off the script as it is AI, so many people tend to try and make them say the wrong things, for example DPD’s AI chatbot recently went rogue against its creator.
DPD AI Bot Goes Rogue
Twitter user Ashley Beauchamp shared their experience with DPDs AI chatbot and managed to turn it against DPD. The chatbot is completely useless at answering any queries, which is about standard for any courier, but one thing it was good at is creating a poem about how terrible DPD is as seen in the shared video. Beauchamp then managed to get the bot to go against its “I’m not allowed to swear” rules by simply asking it to swear in its future answers to me. The bot is then asked to recommend a better delivery firm all whilst exaggerating its hatred of DPD.
A response from DPD to this AI balls-up (Via Techreport) blames the issues upon an update that the chatbot received just a day before the incident which makes perfect sense. It’s a risk you take with using an AI chatbot as opposed to a human, but using DPD is also a risk you take, although in the UK they’re probably the best option compared to the competition. The parcel may not make it to you, but at least they don’t just toss it on the roof or into the neighbours window.
