Irina Raicu is the director of the Internet Ethics program (@IEthics) at the Markkula Center for Applied Ethics. Views are her own.
The chatbots are here, and are getting integrated into more and more services.
Last week, many who had been unaware of the existence of a platform called Discord became aware of it while learning about the massive leak of classified documents via one of the servers that is part of Discord’s network. Now, different coverage focuses on the platform’s integration of OpenAI technology into its chatbot, Clyde.
TechCrunch reports that users have been able to “trick” Clyde into giving them instructions for making things like napalm and meth. In a blog post describing Clyde, Discord warns “that even with its safeguards, Clyde is ‘experimental and might respond with content or other information that could be considered biased, misleading, harmful, or inaccurate.’”
On a related note, a Bloomberg article about concerns among Google employees who thought that their own chatbot was not ready for broad deployment states that “Google’s leaders decided that as long as it called new products ‘experiments,’ the public might forgive their shortcomings, [some] employees said.”
There seems to be a trend developing—a hope that the word “experimental” will serve as some kind of talisman. We’ll likely soon find out whether calling a product “experimental” somehow protects its producers and deployers from accountability and liability if the product is released into society--or, as lawyers might put it, into the “stream of commerce”—especially if that product is then quickly integrated into many others.
So many “experimental” products released in a rush, into our information streams, our classrooms, some of our work spaces, our difficult conversations. A vast experiment to determine how quickly regulators (and courts) can move, how quickly the media can spread hype or fear, how quickly educators and artists and others can adapt, how quickly we all fall for a simulacrum of human interaction.
Speaking of interactions, there are also media reports about scientists working to “reincarnate” the wooly mammoth. I’m imagining the accompanying blog post now:
Note that these mammoths are experimental. Testing them in labs or in zoos would not have given us an accurate, complete picture of how they will interact with other living creatures in the complex environments of the real world, so we decided to release them into a variety of locations, including city neighborhoods, agricultural zones, national parks, porcelain shops, etc. We look forward to learning a lot from public feedback, which will help us in our ongoing reincarnation efforts.
For further reading: some resources for organizations that might be interested in experimenting with ethics training in technology practice: https://www.scu.edu/ethics-in-technology-practice/
"202003 Woolly mammoth" by DataBase Center for Life Science (DBCLS) is licensed under CC BY 4.0.