talking to AI --9/17/25

17 Sep 2025

talking to AI --9/17/25

0 COMMENTS

Today's selection -- from Co-Intelligence by Ethan Mollick. Wharton professor Ethan Mollick explains how we should treat Al like a person (but tell it what kind of person it is).

“I'm about to commit a sin. And not just once, but many, many times. For the rest of this book, I am going to anthropomorphize AI. That means I am going to stop writing that an ‘Al “thinks” something’ and instead just write that ‘AI thinks something.’ The missing quotation marks may seem like a subtle distinction, but it is an important one. Many experts are very nervous about anthropomorphizing AI, and for good reason.

“Anthropomorphism is the act of ascribing human characteristics to something that is nonhuman. We're prone to this: we see faces in the clouds, give motivations to the weather, and hold conversations with our pets. It's no surprise, then, that we're tempted to anthropomorphize artificial intelligence, especially since talking to LLMs feels so much like talking to a person. Even the developers and researchers who design these systems can fall into the trap of using humanlike terms to describe their creations. We say that these complex algorithms and computations ‘understand,’ ‘learn,’ and even ‘feel,’ creating a sense of familiarity and relatability but also, potentially, confusion and misunderstanding.

“This may seem like a silly thing to worry about. After all, it is just a harmless quirk of human psychology, a testament to our ability to empathize and connect. But a lot of researchers are deeply concerned about the implications of casually acting as if AI is a human, both ethically and epistemologically. As researchers Gary Marcus and Sasha Luccioni warn, ‘The more false agency people ascribe to them, the more they can be exploited.’ Consider the humanlike interface of Als like Claude or Siri, or social robots and therapeutic Als explicitly designed to create the illusion of a sympathetic human on the other side. While anthropomorphism might serve a useful purpose in the short term, it raises ethical questions about deception and emotional manipulation. Are we being ‘fooled’ into believing these machines share our feelings? And could this illusion lead us to disclose personal information to these machines, not realizing that we are sharing with corporations or remote operators?

“Treating AI like a person can create unrealistic expectations, false trust, or unwarranted fear among the public, policy makers, and even researchers themselves. It can obscure the true nature of AI as software, leading to misconceptions about its capabilities. It can even influence how we interact with AI systems, affecting our well-being and social relationships.

“Thus, in the following chapters, when I say an AI ‘thinks,’ ‘learns,’ ‘understands,’ ‘decides,’ or ‘feels,’ please remember that I'm speaking metaphorically. AI systems don't have a consciousness, emotions, a sense of self, or physical sensations. But I will pretend that they do for one simple, and one complex, reason. The simple reason is narrative; it's difficult to tell a story about things and much easier to tell a story about beings. The more complex reason: as imperfect as the analogy is, working with AI is easiest if you think of it like an alien person rather than a human-built machine.

“So let's start sinning. Imagine your AI collaborator as an infinitely fast intern, eager to please but prone to bending the truth. Despite our history of thinking about AI as unfeeling, logical robots, LLMs act more like humans. They can be creative, witty, and persuasive, but they can also be evasive and make up plausible, but wrong, information when pressed to give an answer. They are not experts in any domain, but they can mimic the language and style of experts in ways that can be either helpful or misleading. They are unaware of the real world but can generate plausible scenarios and stories based on common sense and patterns. They are not your friends (for now) but can adapt to your preferences and personality by learning from your feedback and interactions. They even seem to respond to emotional manipulation, with researchers documenting that LLMs produce better answers if you tell them ‘this is important to my career’ as part of your prompt. They are, in short, suggestible and even gullible.

“To make the most of this relationship, you must establish a clear and specific AI persona, defining who the Al is and what. problems it should tackle. Remember that LLMs work by pre; dieting the next word, or part of a word, that would come after your prompt. Then they continue to add language from there, again predicting which word will come next. So the default output of many of these models can sound very generic, since they tend to follow similar patterns common in the written documents the AI was trained on. By Breaking the pattern, you can get much more useful and interesting outputs. The easiest way to do that is to provide context and constraints. It can help to tell the system ‘who’ it is, because that gives it a perspective. Telling it to act as a teacher of MBA students will result in a different output than if you ask it to act as a circus clown. This isn't magical—you can't say Act as Bill Gates and get better business advice-but it can help make the tone and direction appropriate for your purpose.

“Let me give you some examples of how defining an AI persona can improve your results. Suppose you're a marketing director and you want to use an LLM to generate some catchy slogans for your new product, a smartwatch that tracks your health and fitness. If you simply ask the LLM, Generate some slogans for my product, without any context or constraints, you might get something like this:

"- The smartwatch that keeps you smart
- Your health and fitness partner
- The ultimate device for your lifestyle
- Stay connected and healthy with our smartwatch

“These are not very original or memorable. They sound like generic marketing phrases that could apply to any smartwatch or wearable device. They don't capture what makes your product unique or appealing. Now suppose you give the LLM some context and constraints by telling it who it is and what it should do. For example, you could say, Act as a witty comedian and generate some slogans for my product that make people laugh. Then you might get something like: The ultimate device for lazy people who want to look fit.Or: Why hire a personal trainer when your wrist can nag you for free? (Though, as you can probably see, most Ais prefer to stay in Dad Joke territory.)

"Of course, you don't have to have the AI act as a comedian if that's not your style or goal. You could also ask it to act as an expert, a friend, a critic, a storyteller, or any other role that suits your purpose. The key is to give the LLM some guidance and direction on how to generate outputs that match your expectations and needs, to put it in the right "headspace" to give you interesting and unique answers. Research has shown that asking the AI to conform to different personas results in differ ent, and often better, answers. But it isn't always clear what personas work best, and LLMs may even subtly adapt their persona to your questioning technique, providing less accurate answers to people who seem less experienced, so experimentation is key.

“Once you give it a persona, you can work with it as you would another person or an intern. I witnessed the value of this approach in action when I assigned my students to "cheat" by using an AI to generate a five-paragraph essay on a relevant topic. At first, the students gave simple and vague prompts, resulting in mediocre essays. But as they tried different strategies, the quality of the Al's output improved significantly. One very effective strategy that emerged from the class was treating the AI as a coeditor, engaging in a back-and-forth, conversational process. Students produced impressive essays that far exceeded their initial attempts by constantly refining and re directing the AI.

“Remember, your AI intern, though incredibly fast and knowledgeable, is not flawless. It's crucial to keep a critical eye on and treat the AI as a tool that works for you. By defining its persona, engaging in a collaborative editing process, and continually providing guidance, you can take advantage of AI as a form of collaborative co-intelligence."