Impromptu: Amplifying Our Humanity Through AI The first is that AI tools capable of communicating in human- like ways are by no means inherently deceptive. It’s all about intention. A customer-service chatbot that clearly conveys its status as a non-human entity is in fact behaving in an extremely authentic manner. The second is that while AI tools can potentially cause confu- sion when used deceptively, they will also create great value when used transparently and straightforwardly. Everyone has entered the chat It took less than a week for ChatGPT to show how much people actually like conversing with chatbots when those chat- bots possess human-like fluency instead of a narrow range of canned responses. So imagine when every airline, shipping service, online retailer, and government agency in the world begins to deploy chatbots as capable of answering your ques- tions as GPT-4. Now, to extend this further—and to bring this line of thought back to social media—imagine having extended conversations with chatbot versions of MrBeast, Michelle Obama, Benjamin Franklin, Marge Simpson, the world’s best high school math tutor, an expert on B2B marketing, etc. I’m not saying this world is coming next month, or even next year. But I do think that LLM chatbots of this type will even- tually join the ranks of books, podcasts, instructional videos, music albums, and other media formats that individuals use to syndicate their thoughts, values, personality, and creativity across time and space. Of course, this isn’t an endeavor without risk, at least as long as LLMs continue to produce hallucinations and other unin- 104