ChatGPT is currently the talk of the town. The garden beekeeper tried out the "intelligent" chatbot with the fictitious task of generating a text for an advertising mail:

 

[Input] Write an advertisement email about Garden-Honey. The email should say that there is still honey in stock but whoever wants to buy some honey should be quick.

 

 

[Output] Dear valued customers,

 

We are pleased to announce that we still have a limited supply of our delicious Garden-Honey available for purchase. If you have been searching for the perfect honey to add to your pantry, now is the time to act!

 

Our Garden-Honey is made from the nectar of the finest flowers and plants in our gardens, resulting in a unique and flavorful taste that you won't find anywhere else. It is the perfect addition to any recipe, from breakfast toast to baked goods.

 

However, we want to warn you that our honey is in high demand and our supplies are running low. If you want to get your hands on some of our Garden-Honey, we suggest placing your order as soon as possible to avoid disappointment.

 

Thank you for choosing Garden-Honey. We look forward to hearing from you soon.

 

Sincerely,

[Your Name]

[Your Company]

 

Garden Beekeeper finds this result surprisingly good. At second glance, however, it is rather astonishingly eloquently babbled, and whether this information corresponds to the truth or is a lie can hardly be ascertained and is thus left to the reader's power of judgment. Because this AI cannot know the truth. And that is indeed a bit creepy....

 

Background info e.g. here: https://www.spektrum.de/news/maschinelles-lernen-chatgpt-wird-immer-plappern/2090727