Lecture 28: AI And Humans

class: center, middle

# Artificial Intelligence

## AI and Humans

---

# Humans and Computers

.left-column[
Humans are:

- Creative

- Social

- Critical

- Imprecise

]

.right-column[
Computer are:

- (Mostly) deterministic

- Uncaring

- Uncritical

- Precise
]

---

# A little bit of communication theory

* Human communication follows certain patterns
 
 * The goal of a conversation is usually mutual understanding by the participants
 
 * Note that several situations cause violations of this idea: Jokes, surprise, deception, etc.
 
 * However, general conversations follow the *cooperative principle*

---

# Grice's Maxims of Communication

* The cooperative principle was divided into four maxims by H.P. Grice 
  
     - The Maxim of Quality
     
     - The Maxim of Quantity 
     
     - The Maxim of Relation
     
     - The Maxim of Manner
     
---

# Visual Chatbot: Question

---

# Why?

* The training data are image descriptions

* But consider Grice's Maxim of Relation!

* A human will not (usually) give "There are no giraffes in this picture" as a description

* Only when there **are** giraffes will the human say something

* These language models are also **not** "intelligent": They have no sense of reality

* Play with the [Visual Chatbot](http://demo.visualdialog.org/)

---

# And therefore

I asked GPT-3:

> *How many plosarei are in a ocpolisa?*

<br/>
It answered:

> There are about as many as there are people, but they all act like one big unit now.

---

# Text Generation

* The problem with text generation is that it is ill-defined

* There is no "scoring" function, nothing to "optimize"

* Instead, there are many different shades of "correct"

* Grammar might be perfect, context might be perfect, but the semantics are easy to mess up

---

# Creativity

* So, how do you judge "makes sense"?

* Or if you generate pictures, stories, movies, music, etc., how do you determine if your generator is "good"?

* It depends!

---

# Good Stories

* Let's consider the problem of generating stories

* What makes a story "good"?

* Depends on the goal!

* Possible goals: Believable characters, memorable events, uniqueness, conveying a particular message, ...

---

# Specialized measures

* The more concrete we can get with our goal, the more feasible it becomes to measure

* For example, say we want a story that conveys some information

* We could model this as a query to the mental state of the reader after reading the story

---

class: medium

# QUEST

* QUEST is a cognitive model of question answering

* Given a story, a graph representation of the information contained in it is built

* Then there are operations to query this graph

* It has been shown (Cardona-Rivera et al., 2016) that this knowledge structure can predict a human reader's answers

* Of course, the model is not complete and does not pick up on all nuances

---

# Novelty

* Humans are also often drawn to novelty

* If you tell someone that a story, movie, picture, etc. was generated by a computer it may automatically be "more interesting"

* This is not necessarily bad, but it may affect the perception

---

# Social Simulations

* Social situations are complicated for AI agents

* There are several models for aspects of social behavior

* For example, one may model opinion exchange

* These models can then be validated piece by piece

---

# Self-Analysis

* The nice thing about computers it that they do what you tell them

* The problem with computers is that they do exactly what you tell them

* Who is responsible if something goes wrong?

* How can you tell why a computer did something?

---

# Explainable AI

* There is a push towards explainability

* Which inputs affected the output?

* How does the output change if the inputs change?

* This is a difficult problem for some approaches (remember Neural Networks?)

---

# Communication

* Another challenge with communication is ambiguity and context

* Recall Grice: Maxim of Manner

* But people are ambiguous all the time, because they know the *context*

* Words alone rarely capture the entirety of communication

---

# Literal Interpretation

---

# Precision

* Computers have the advantage of being precise, especially with numbers

* Humans are not

* This sometimes leads to friction

---

# Precision

---

# Precision

---

# Less Precision

---

# Implicit Speech Acts

A: "Can you pass me the salt?"

B: "Yes, I am capable of this action"

---

# Implicit Speech Acts

A: "Can you pass me the salt?"

B: "Yes, I am capable of this action"

<br/>

A: "It is pretty hot in here."

B, standing by the thermostat:

---

# Implicit Speech Acts

* **Context** is important

* In physical therapy, asking if a patient is physically capable of lifting a salt shaker may happen

* But usually such questions are not meant to evoke a literal *answer*

* Sometimes action requests aren't even questions

---

# Subtle Communication

<video width="720" height="404" controls>
  <source src="/CS4200/assets/img/cut.mp4" type="video/mp4">
  <source src="/CS4200/assets/img/cut.avi" type="video/avi">
  Your browser does not support the video tag.
</video>

---

# Subtle Communication

* What does a viewer actually look at?

* What do they remember?

* What do they notice while not looking at it?

* How do you draw attention to clues?

---

# More Subtlety

* Real Life: "Beyond Reasonable Doubt"

* Detective Stories: "Beyond Casual Doubt"

* The detective is "logical"

* But are they really?

---

# Variety

* The "appeal" of detective stories comes not only from the structure

* Each "episode" or "story" plays in a wildly different "world"

* Elementary alone had: Mathematician, Art Dealers, Conspiracy Theorists, Base Jumpers, etc.

* For this, the AI system needs to "understand" a variety of contexts

---

# Why?

* All this makes Detective Stories a great application for AI!

* There is a foundation in logic ... but not too much

* There is a need for understanding the world ... to an extend

* There is a challenge of communication

---

# References

* [AI Dungeon](https://play.aidungeon.io/)

* [Question Answering in the Context of Stories Generated by Computers](http://www.rogel.io/content/2-publications/cardona-rivera2016question.pdf)

* [Opinionated Virtual Characters](https://www.aaai.org/ojs/index.php/AIIDE/article/download/5232/5088)

* [Murder Mysteries: The White Whale of Narrative Generation?](https://ojs.aaai.org//index.php/AIIDE/article/view/7432)