From bda04b63bd79cb4f4cd182b6da2124e6b8a69a09 Mon Sep 17 00:00:00 2001 From: Candace Savonen Date: Fri, 15 Dec 2023 09:56:08 -0500 Subject: [PATCH 1/8] Set up quiz files fix numbering --- ...e.Rmd => 04c-AI_Policy-how_to_navigate.Rmd | 0 ...ta_laws.Rmd => 04d-AI_Policy-data_laws.Rmd | 0 quizzes/04-AI_Policy_overall_quiz.md | 7 +++ .../04b-AI_Policy-ai_regs_and_laws_quiz.md | 7 +++ quizzes/04c-AI_Policy-how_to_navigate_quiz.md | 7 +++ quizzes/04d-AI_Policy-data_laws_quiz.md | 6 ++ ...-AI_Policy-creating_your_ai_policy_quiz.md | 7 +++ ...AI_Policy-education_and_compliance_quiz.md | 7 +++ quizzes/quiz_ch1.md | 61 ------------------- 9 files changed, 41 insertions(+), 61 deletions(-) rename 04b-AI_Policy-how_to_navigate.Rmd => 04c-AI_Policy-how_to_navigate.Rmd (100%) rename 04c-AI_Policy-data_laws.Rmd => 04d-AI_Policy-data_laws.Rmd (100%) create mode 100644 quizzes/04-AI_Policy_overall_quiz.md create mode 100644 quizzes/04b-AI_Policy-ai_regs_and_laws_quiz.md create mode 100644 quizzes/04c-AI_Policy-how_to_navigate_quiz.md create mode 100644 quizzes/04d-AI_Policy-data_laws_quiz.md create mode 100644 quizzes/04e-AI_Policy-creating_your_ai_policy_quiz.md create mode 100644 quizzes/04f-AI_Policy-education_and_compliance_quiz.md delete mode 100644 quizzes/quiz_ch1.md diff --git a/04b-AI_Policy-how_to_navigate.Rmd b/04c-AI_Policy-how_to_navigate.Rmd similarity index 100% rename from 04b-AI_Policy-how_to_navigate.Rmd rename to 04c-AI_Policy-how_to_navigate.Rmd diff --git a/04c-AI_Policy-data_laws.Rmd b/04d-AI_Policy-data_laws.Rmd similarity index 100% rename from 04c-AI_Policy-data_laws.Rmd rename to 04d-AI_Policy-data_laws.Rmd diff --git a/quizzes/04-AI_Policy_overall_quiz.md b/quizzes/04-AI_Policy_overall_quiz.md new file mode 100644 index 00000000..56f60f09 --- /dev/null +++ b/quizzes/04-AI_Policy_overall_quiz.md @@ -0,0 +1,7 @@ + +{quiz, id: policy_overall, attempts: 10} + +## AI Policy Overall quiz + + +{/quiz} diff --git a/quizzes/04b-AI_Policy-ai_regs_and_laws_quiz.md b/quizzes/04b-AI_Policy-ai_regs_and_laws_quiz.md new file mode 100644 index 00000000..e8a941da --- /dev/null +++ b/quizzes/04b-AI_Policy-ai_regs_and_laws_quiz.md @@ -0,0 +1,7 @@ + +{quiz, id: regs_laws, attempts: 10} + +## AI Regulations and Laws quiz + + +{/quiz} diff --git a/quizzes/04c-AI_Policy-how_to_navigate_quiz.md b/quizzes/04c-AI_Policy-how_to_navigate_quiz.md new file mode 100644 index 00000000..cadbfa9e --- /dev/null +++ b/quizzes/04c-AI_Policy-how_to_navigate_quiz.md @@ -0,0 +1,7 @@ + +{quiz, id: how_to_navigate, attempts: 10} + +## How to Navigate quiz + + +{/quiz} diff --git a/quizzes/04d-AI_Policy-data_laws_quiz.md b/quizzes/04d-AI_Policy-data_laws_quiz.md new file mode 100644 index 00000000..ff955b1e --- /dev/null +++ b/quizzes/04d-AI_Policy-data_laws_quiz.md @@ -0,0 +1,6 @@ + +{quiz, id: data_laws, attempts: 10} + +## Data Laws Quiz + +{/quiz} diff --git a/quizzes/04e-AI_Policy-creating_your_ai_policy_quiz.md b/quizzes/04e-AI_Policy-creating_your_ai_policy_quiz.md new file mode 100644 index 00000000..feeb2679 --- /dev/null +++ b/quizzes/04e-AI_Policy-creating_your_ai_policy_quiz.md @@ -0,0 +1,7 @@ + +{quiz, id: creating_policy, attempts: 10} + +## Creating AI Policy quiz + + +{/quiz} diff --git a/quizzes/04f-AI_Policy-education_and_compliance_quiz.md b/quizzes/04f-AI_Policy-education_and_compliance_quiz.md new file mode 100644 index 00000000..6f0af0ba --- /dev/null +++ b/quizzes/04f-AI_Policy-education_and_compliance_quiz.md @@ -0,0 +1,7 @@ + +{quiz, id: compliance, attempts: 10} + +## Policy Education and Compliance quiz + + +{/quiz} diff --git a/quizzes/quiz_ch1.md b/quizzes/quiz_ch1.md deleted file mode 100644 index df3fe8d6..00000000 --- a/quizzes/quiz_ch1.md +++ /dev/null @@ -1,61 +0,0 @@ - -{quiz, id: quiz_name_here, attempts: 10} - -## Template quiz - -Put any other instructions your quiz takers need to know here like: Choose the best answer. - -{choose-answers: 4} -? First question to ask goes here. (Note- you need a question mark at end like this. Just one is required if using a question mark in your question field)? - -C) One correct answer here marked with a "C" -C) A second correct answer here -m) Mandatory incorrect answers have an "m" -m) A second mandatory incorrect answer -o) An optional incorrect answer here marked with an "o" -o) A second optional incorrect answer here - -{choose-answers: 4} -? Question example with just a question mark? - -C) One correct answer here marked with a "C" -C) A second correct answer here -m) Mandatory incorrect answers have an "m" -m) A second mandatory incorrect answer -o) An optional incorrect answer here marked with an "o" -o) A second optional incorrect answer here - -{choose-answers: 4} -? Second question to ask goes here? - -C) A correct answer here! -m) Mandatory incorrect answers have an "m" -m) A second mandatory incorrect answer -o) An optional incorrect answer here marked with an "o" -o) A second optional incorrect answer here - - -{choose-answers: 4} -? A more complicated example. Note the question mark at the end of the options! Which of the following are correct -1. An option described -2. A second option -3. A third option -4. A fourth option -5. A fifth option ? - -C) All of the examples listed except 5 -C) All of the listed examples except 5 -m) 1, 3, and 5 -m) 1, 2, and 3 -o) All of the examples except 1 and 5 -o) All of the examples listed - -{random-choice-order: false} -? A question in which the order of choices is important? - -a) The possible responses should be labeled a, b, c, etc -b) The correct answer has a capital letter, in this case "D" is correct -c) both a and b -D) All of the above will always be last choice in this question - -{/quiz} From 37b0dcb4927945523957e63ccb783cab58c5d1b5 Mon Sep 17 00:00:00 2001 From: Candace Savonen Date: Mon, 18 Dec 2023 13:55:37 -0500 Subject: [PATCH 2/8] Adding quizzes --- quizzes/04b-AI_Policy-ai_regs_and_laws_quiz.md | 17 +++++++++++++++++ quizzes/04c-AI_Policy-how_to_navigate_quiz.md | 16 ++++++++++++++-- 2 files changed, 31 insertions(+), 2 deletions(-) diff --git a/quizzes/04b-AI_Policy-ai_regs_and_laws_quiz.md b/quizzes/04b-AI_Policy-ai_regs_and_laws_quiz.md index e8a941da..c37ae9a2 100644 --- a/quizzes/04b-AI_Policy-ai_regs_and_laws_quiz.md +++ b/quizzes/04b-AI_Policy-ai_regs_and_laws_quiz.md @@ -3,5 +3,22 @@ ## AI Regulations and Laws quiz +? What is the main goal of the EU's AI Act according to policymakers? +a) To ban all AI systems +B) To limit the capacity of AI systems to cause harm +c) To standardize AI systems across Europe +d) To tax companies using AI systems + +? Which of the following must developers provide for general purpose AI systems like GPT under the EU AI Act? +a) A list of potential biases +b) Documentation about the training data +c) Monthly accuracy reports +d) Cybersecurity audit reports + +? Which of these statements is true about the EU AI act? +A) There are exceptions to the facial scraping ban that allow law enforcement and intelligence agencies to use AI applications for facial recognition purposes. +b) AI software for manipulative strategies like deepfakes and automated disinformation campaigns are allowed under certain conditions. +c) Companies do not need to disclose to customers if they are using AI tools. + {/quiz} diff --git a/quizzes/04c-AI_Policy-how_to_navigate_quiz.md b/quizzes/04c-AI_Policy-how_to_navigate_quiz.md index cadbfa9e..310b8207 100644 --- a/quizzes/04c-AI_Policy-how_to_navigate_quiz.md +++ b/quizzes/04c-AI_Policy-how_to_navigate_quiz.md @@ -1,7 +1,19 @@ -{quiz, id: how_to_navigate, attempts: 10} +{quiz, id: building_a_team, attempts: 10} -## How to Navigate quiz +## Building a team to guide your AI use quiz +? True or false - The author recommends having only technical experts like data scientists and AI researchers on an AI oversight team. +a) True +B) False + +? Which of the following were not individuals listed as part of an AI team? + +a) Legal council +b) Policy and governance analysts +c) Data protection officers +d) Ethicists and oversight board members +e) Marketers +f) Trainers and educators {/quiz} From de7041ee726bc7e418bd4921e9815253159873d5 Mon Sep 17 00:00:00 2001 From: Candace Savonen Date: Mon, 18 Dec 2023 14:24:02 -0500 Subject: [PATCH 3/8] Add overall quiz --- quizzes/04-AI_Policy_overall_quiz.md | 43 +++++++++++++++++++ quizzes/04d-AI_Policy-data_laws_quiz.md | 17 ++++++++ ...-AI_Policy-creating_your_ai_policy_quiz.md | 17 ++++++++ ...AI_Policy-education_and_compliance_quiz.md | 7 --- 4 files changed, 77 insertions(+), 7 deletions(-) delete mode 100644 quizzes/04f-AI_Policy-education_and_compliance_quiz.md diff --git a/quizzes/04-AI_Policy_overall_quiz.md b/quizzes/04-AI_Policy_overall_quiz.md index 56f60f09..36339320 100644 --- a/quizzes/04-AI_Policy_overall_quiz.md +++ b/quizzes/04-AI_Policy_overall_quiz.md @@ -4,4 +4,47 @@ ## AI Policy Overall quiz +? What are examples of user agreements? +a) Commercial agreements +b) Enterprise agreements +c) Intellectual agreements +d) Both A and C +E) Both A and B +f) All of the above + +? Which of the following were NOT a topic we discussed that may be a part of an AI policy? +a) Governance and Oversight +b) Education and Training +C) Pricing points for customers +d) Feedback and Review + +? Which of the following are legitimate pieces of advice your should adhere to regarding AI policy? +a) Always consult your legal council about the AI regulations that apply to you. +b) It is not enough to build a policy, support systems need to be in place to support individuals in their ability to follow that policy. +c) You should seek advice from a variety of experts and sources when building a comprehensive AI policy +d) A and B are true +e) B and C are true +F) All of the above are true + + +? What is one concern around the use of copyrighted data to train AI models? +a) The models may produce biased results +B) The models may violate intellectual property rights +c) The models may be insecure +d) The models will use too much computer processing power + +? What does the text recommend to allow an AI policy to stay relevant as technology changes quickly? +a) Update it daily +b) Make it a short-term policy only +C) Create it as an ongoing living document +d) Don't create a policy at all + +? An effective AI policy... + +a) Is agile and readily adjusted as the landscape of technology changes +b) Is accompanied by training, education and infrastructure +c) Has been created by a team of experts weighing in +d) Discusses how bias and discrimination is being monitored +E) All of the above should be included with an effective AI policy!x + {/quiz} diff --git a/quizzes/04d-AI_Policy-data_laws_quiz.md b/quizzes/04d-AI_Policy-data_laws_quiz.md index ff955b1e..18f01be6 100644 --- a/quizzes/04d-AI_Policy-data_laws_quiz.md +++ b/quizzes/04d-AI_Policy-data_laws_quiz.md @@ -3,4 +3,21 @@ ## Data Laws Quiz +? What does copyright protect? +a) Ideas for new inventions +B) The exclusive right to publish or reproduce artistic material +c) Software code +d) Scientific discoveries + +? What type of information is covered under special data privacy protections? +a) Social media posts +B) Personal health information +c) Public government reports +d) News articles + +? Who is most likely legally liable if an AI system was inherently flawed? +a) The end user of the AI system +b) The developer of the AI algorithm +C) The designer or manufacturer of the system + {/quiz} diff --git a/quizzes/04e-AI_Policy-creating_your_ai_policy_quiz.md b/quizzes/04e-AI_Policy-creating_your_ai_policy_quiz.md index feeb2679..e8356607 100644 --- a/quizzes/04e-AI_Policy-creating_your_ai_policy_quiz.md +++ b/quizzes/04e-AI_Policy-creating_your_ai_policy_quiz.md @@ -3,5 +3,22 @@ ## Creating AI Policy quiz +? According to the text, what should an AI policy support system include in addition to the policy itself? +a) Guidelines on best practices only +b) Infrastructure to handle AI only +c) Training materials only +D) Guidelines, infrastructure, and training materials + +? Who should be involved in creating a comprehensive AI policy? +a) Chief Data Officer's office only +b) IT department only +c) Legal department only +D) Representatives from legal, IT, compliance, governance, diversity offices, etc. + +? What should training on an AI policy focus on? +a) Explaining every detail of the policy +B) Outlining the key points to consider and detailing who they should consult for guidance +c) Punishments for non-compliance +d) Using scare tactics about AI going wrong {/quiz} diff --git a/quizzes/04f-AI_Policy-education_and_compliance_quiz.md b/quizzes/04f-AI_Policy-education_and_compliance_quiz.md deleted file mode 100644 index 6f0af0ba..00000000 --- a/quizzes/04f-AI_Policy-education_and_compliance_quiz.md +++ /dev/null @@ -1,7 +0,0 @@ - -{quiz, id: compliance, attempts: 10} - -## Policy Education and Compliance quiz - - -{/quiz} From 5f7190aeacfabcb1e6df35ba129888cf609e0e4d Mon Sep 17 00:00:00 2001 From: Candace Savonen Date: Mon, 18 Dec 2023 14:25:22 -0500 Subject: [PATCH 4/8] Add spelling to dictionary --- resources/dictionary.txt | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/resources/dictionary.txt b/resources/dictionary.txt index 017cadac..bcc414d5 100644 --- a/resources/dictionary.txt +++ b/resources/dictionary.txt @@ -102,6 +102,7 @@ NHGRI NLP NPS OpenAI +OctoML ottrpal Pandoc Perelman @@ -113,6 +114,7 @@ premade Premade PrivateAI programmatically +regs relIabiliTy reproducibility rollout @@ -120,8 +122,9 @@ Samsung scORe skillsetes skillsets -timeframes +timeframes transformative +Ubers UE UE5 under-diagnosed From dffaf20e72471a506ac6a7e13a2b5d20eec487d2 Mon Sep 17 00:00:00 2001 From: Candace Savonen Date: Mon, 18 Dec 2023 14:27:43 -0500 Subject: [PATCH 5/8] Fix quizzes --- quizzes/04-AI_Policy_overall_quiz.md | 1 - quizzes/04c-AI_Policy-how_to_navigate_quiz.md | 2 +- 2 files changed, 1 insertion(+), 2 deletions(-) diff --git a/quizzes/04-AI_Policy_overall_quiz.md b/quizzes/04-AI_Policy_overall_quiz.md index 36339320..f7c410a6 100644 --- a/quizzes/04-AI_Policy_overall_quiz.md +++ b/quizzes/04-AI_Policy_overall_quiz.md @@ -3,7 +3,6 @@ ## AI Policy Overall quiz - ? What are examples of user agreements? a) Commercial agreements b) Enterprise agreements diff --git a/quizzes/04c-AI_Policy-how_to_navigate_quiz.md b/quizzes/04c-AI_Policy-how_to_navigate_quiz.md index 310b8207..784a252d 100644 --- a/quizzes/04c-AI_Policy-how_to_navigate_quiz.md +++ b/quizzes/04c-AI_Policy-how_to_navigate_quiz.md @@ -13,7 +13,7 @@ a) Legal council b) Policy and governance analysts c) Data protection officers d) Ethicists and oversight board members -e) Marketers +E) Marketers f) Trainers and educators {/quiz} From 03f93d89b3068db66aa97d6433c0cad56eb36d2e Mon Sep 17 00:00:00 2001 From: Candace Savonen Date: Mon, 18 Dec 2023 14:30:37 -0500 Subject: [PATCH 6/8] Fix bookdown. Add coursera yamls --- _bookdown.yml | 4 +- .../04-AI_Policy_overall_quiz.md.yml | 89 +++++++++++++++++++ ...04b-AI_Policy-ai_regs_and_laws_quiz.md.yml | 37 ++++++++ .../04c-AI_Policy-how_to_navigate_quiz.md.yml | 27 ++++++ .../04d-AI_Policy-data_laws_quiz.md.yml | 37 ++++++++ ...Policy-creating_your_ai_policy_quiz.md.yml | 39 ++++++++ 6 files changed, 231 insertions(+), 2 deletions(-) create mode 100644 coursera_quizzes/04-AI_Policy_overall_quiz.md.yml create mode 100644 coursera_quizzes/04b-AI_Policy-ai_regs_and_laws_quiz.md.yml create mode 100644 coursera_quizzes/04c-AI_Policy-how_to_navigate_quiz.md.yml create mode 100644 coursera_quizzes/04d-AI_Policy-data_laws_quiz.md.yml create mode 100644 coursera_quizzes/04e-AI_Policy-creating_your_ai_policy_quiz.md.yml diff --git a/_bookdown.yml b/_bookdown.yml index 6b92166b..597c692b 100644 --- a/_bookdown.yml +++ b/_bookdown.yml @@ -25,9 +25,9 @@ rmd_files: ["index.Rmd", "03f-Determining-AI-Needs-interface.Rmd", "03g-Determining-AI-Needs-evaluation.Rmd", "04a-AI_Policy-intro.Rmd", - "04b-AI_Policy-how_to_navigate.Rmd", "04b-AI_Policy-ai_regs_and_laws.Rmd", - "04c-AI_Policy-data_laws.Rmd", + "04c-AI_Policy-how_to_navigate.Rmd", + "04d-AI_Policy-data_laws.Rmd", "04e-AI_Policy-creating_your_ai_policy.Rmd", "About.Rmd", "References.Rmd"] diff --git a/coursera_quizzes/04-AI_Policy_overall_quiz.md.yml b/coursera_quizzes/04-AI_Policy_overall_quiz.md.yml new file mode 100644 index 00000000..c438f503 --- /dev/null +++ b/coursera_quizzes/04-AI_Policy_overall_quiz.md.yml @@ -0,0 +1,89 @@ + +- typeName: multipleChoice + prompt: What are examples of user agreements? + shuffleOptions: true + options: + - answer: Commercial agreements + isCorrect: false + - answer: Enterprise agreements + isCorrect: false + - answer: Intellectual agreements + isCorrect: false + - answer: Both A and C + isCorrect: false + - answer: Both A and B + isCorrect: true + - answer: All of the above + isCorrect: false + +- typeName: multipleChoice + prompt: Which of the following were NOT a topic we discussed that may be a part of an AI policy? + shuffleOptions: true + options: + - answer: Governance and Oversight + isCorrect: false + - answer: Education and Training + isCorrect: false + - answer: Pricing points for customers + isCorrect: true + - answer: Feedback and Review + isCorrect: false + +- typeName: multipleChoice + prompt: Which of the following are legitimate pieces of advice your should adhere to regarding AI policy? + shuffleOptions: true + options: + - answer: Always consult your legal council about the AI regulations that apply to you. + isCorrect: false + - answer: It is not enough to build a policy, support systems need to be in place to support individuals in their ability to follow that policy. + isCorrect: false + - answer: You should seek advice from a variety of experts and sources when building a comprehensive AI policy + isCorrect: false + - answer: A and B are true + isCorrect: false + - answer: B and C are true + isCorrect: false + - answer: All of the above are true + isCorrect: true + +- typeName: multipleChoice + prompt: What is one concern around the use of copyrighted data to train AI models? + shuffleOptions: true + options: + - answer: The models may produce biased results + isCorrect: false + - answer: The models may violate intellectual property rights + isCorrect: true + - answer: The models may be insecure + isCorrect: false + - answer: The models will use too much computer processing power + isCorrect: false + +- typeName: multipleChoice + prompt: What does the text recommend to allow an AI policy to stay relevant as technology changes quickly? + shuffleOptions: true + options: + - answer: Update it daily + isCorrect: false + - answer: Make it a short-term policy only + isCorrect: false + - answer: Create it as an ongoing living document + isCorrect: true + - answer: Don't create a policy at all + isCorrect: false + +- typeName: multipleChoice + prompt: An effective AI policy... + + shuffleOptions: true + options: + - answer: Is agile and readily adjusted as the landscape of technology changes + isCorrect: false + - answer: Is accompanied by training, education and infrastructure + isCorrect: false + - answer: Has been created by a team of experts weighing in + isCorrect: false + - answer: Discusses how bias and discrimination is being monitored + isCorrect: false + - answer: All of the above should be included with an effective AI policy!x + isCorrect: true diff --git a/coursera_quizzes/04b-AI_Policy-ai_regs_and_laws_quiz.md.yml b/coursera_quizzes/04b-AI_Policy-ai_regs_and_laws_quiz.md.yml new file mode 100644 index 00000000..f7e5c3c9 --- /dev/null +++ b/coursera_quizzes/04b-AI_Policy-ai_regs_and_laws_quiz.md.yml @@ -0,0 +1,37 @@ + +- typeName: multipleChoice + prompt: What is the main goal of the EU's AI Act according to policymakers? + shuffleOptions: true + options: + - answer: To ban all AI systems + isCorrect: false + - answer: To limit the capacity of AI systems to cause harm + isCorrect: true + - answer: To standardize AI systems across Europe + isCorrect: false + - answer: To tax companies using AI systems + isCorrect: false + +- typeName: multipleChoice + prompt: Which of the following must developers provide for general purpose AI systems like GPT under the EU AI Act? + shuffleOptions: true + options: + - answer: A list of potential biases + isCorrect: false + - answer: Documentation about the training data + isCorrect: false + - answer: Monthly accuracy reports + isCorrect: false + - answer: Cybersecurity audit reports + isCorrect: false + +- typeName: multipleChoice + prompt: Which of these statements is true about the EU AI act? + shuffleOptions: true + options: + - answer: There are exceptions to the facial scraping ban that allow law enforcement and intelligence agencies to use AI applications for facial recognition purposes. + isCorrect: true + - answer: AI software for manipulative strategies like deepfakes and automated disinformation campaigns are allowed under certain conditions. + isCorrect: false + - answer: Companies do not need to disclose to customers if they are using AI tools. + isCorrect: false diff --git a/coursera_quizzes/04c-AI_Policy-how_to_navigate_quiz.md.yml b/coursera_quizzes/04c-AI_Policy-how_to_navigate_quiz.md.yml new file mode 100644 index 00000000..bf9a2f15 --- /dev/null +++ b/coursera_quizzes/04c-AI_Policy-how_to_navigate_quiz.md.yml @@ -0,0 +1,27 @@ + +- typeName: multipleChoice + prompt: True or false - The author recommends having only technical experts like data scientists and AI researchers on an AI oversight team. + shuffleOptions: true + options: + - answer: True + isCorrect: false + - answer: False + isCorrect: true + +- typeName: multipleChoice + prompt: Which of the following were not individuals listed as part of an AI team? + + shuffleOptions: true + options: + - answer: Legal council + isCorrect: false + - answer: Policy and governance analysts + isCorrect: false + - answer: Data protection officers + isCorrect: false + - answer: Ethicists and oversight board members + isCorrect: false + - answer: Marketers + isCorrect: true + - answer: Trainers and educators + isCorrect: false diff --git a/coursera_quizzes/04d-AI_Policy-data_laws_quiz.md.yml b/coursera_quizzes/04d-AI_Policy-data_laws_quiz.md.yml new file mode 100644 index 00000000..86b64b6b --- /dev/null +++ b/coursera_quizzes/04d-AI_Policy-data_laws_quiz.md.yml @@ -0,0 +1,37 @@ + +- typeName: multipleChoice + prompt: What does copyright protect? + shuffleOptions: true + options: + - answer: Ideas for new inventions + isCorrect: false + - answer: The exclusive right to publish or reproduce artistic material + isCorrect: true + - answer: Software code + isCorrect: false + - answer: Scientific discoveries + isCorrect: false + +- typeName: multipleChoice + prompt: What type of information is covered under special data privacy protections? + shuffleOptions: true + options: + - answer: Social media posts + isCorrect: false + - answer: Personal health information + isCorrect: true + - answer: Public government reports + isCorrect: false + - answer: News articles + isCorrect: false + +- typeName: multipleChoice + prompt: Who is most likely legally liable if an AI system was inherently flawed? + shuffleOptions: true + options: + - answer: The end user of the AI system + isCorrect: false + - answer: The developer of the AI algorithm + isCorrect: false + - answer: The designer or manufacturer of the system + isCorrect: true diff --git a/coursera_quizzes/04e-AI_Policy-creating_your_ai_policy_quiz.md.yml b/coursera_quizzes/04e-AI_Policy-creating_your_ai_policy_quiz.md.yml new file mode 100644 index 00000000..c3792d86 --- /dev/null +++ b/coursera_quizzes/04e-AI_Policy-creating_your_ai_policy_quiz.md.yml @@ -0,0 +1,39 @@ + +- typeName: multipleChoice + prompt: According to the text, what should an AI policy support system include in addition to the policy itself? + shuffleOptions: true + options: + - answer: Guidelines on best practices only + isCorrect: false + - answer: Infrastructure to handle AI only + isCorrect: false + - answer: Training materials only + isCorrect: false + - answer: Guidelines, infrastructure, and training materials + isCorrect: true + +- typeName: multipleChoice + prompt: Who should be involved in creating a comprehensive AI policy? + shuffleOptions: true + options: + - answer: Chief Data Officer's office only + isCorrect: false + - answer: IT department only + isCorrect: false + - answer: Legal department only + isCorrect: false + - answer: Representatives from legal, IT, compliance, governance, diversity offices, etc. + isCorrect: true + +- typeName: multipleChoice + prompt: What should training on an AI policy focus on? + shuffleOptions: true + options: + - answer: Explaining every detail of the policy + isCorrect: false + - answer: Outlining the key points to consider and detailing who they should consult for guidance + isCorrect: true + - answer: Punishments for non-compliance + isCorrect: false + - answer: Using scare tactics about AI going wrong + isCorrect: false From e3a3d0952a114d33b185fe33500f6ddc62675d0b Mon Sep 17 00:00:00 2001 From: Candace Savonen Date: Mon, 18 Dec 2023 14:56:54 -0500 Subject: [PATCH 7/8] Adding feedback prompts --- .../04-AI_Policy_overall_quiz.md.yml | 13 ++++++++++-- ...04b-AI_Policy-ai_regs_and_laws_quiz.md.yml | 20 ++++++++++--------- .../04c-AI_Policy-how_to_navigate_quiz.md.yml | 8 ++++++++ .../04d-AI_Policy-data_laws_quiz.md.yml | 15 ++++++++++++-- ...Policy-creating_your_ai_policy_quiz.md.yml | 12 +++++++++++ .../04b-AI_Policy-ai_regs_and_laws_quiz.md | 8 +++----- 6 files changed, 58 insertions(+), 18 deletions(-) diff --git a/coursera_quizzes/04-AI_Policy_overall_quiz.md.yml b/coursera_quizzes/04-AI_Policy_overall_quiz.md.yml index c438f503..a848f9da 100644 --- a/coursera_quizzes/04-AI_Policy_overall_quiz.md.yml +++ b/coursera_quizzes/04-AI_Policy_overall_quiz.md.yml @@ -1,20 +1,25 @@ - typeName: multipleChoice prompt: What are examples of user agreements? - shuffleOptions: true options: - answer: Commercial agreements isCorrect: false + feedback: This is one such example of user agreements but there is more - answer: Enterprise agreements isCorrect: false + feedback: This is one such example of user agreements but there is more - answer: Intellectual agreements isCorrect: false + feedback: Intellectual agreements were not something we described as an example of a user agreement - answer: Both A and C isCorrect: false + feedback: Intellectual agreements were not something we described as an example of a user agreement - answer: Both A and B isCorrect: true + feedback: Yes! These are both examples of user agreements - answer: All of the above isCorrect: false + feedback: Intellectual agreements were not something we described as an example of a user agreement - typeName: multipleChoice prompt: Which of the following were NOT a topic we discussed that may be a part of an AI policy? @@ -22,19 +27,23 @@ options: - answer: Governance and Oversight isCorrect: false + feedback: Education and Training was considered a part of effective AI policies - answer: Education and Training isCorrect: false + feedback: Education and Training was considered a part of effective AI policies - answer: Pricing points for customers isCorrect: true + feedback: This was not a part of an AI policy - answer: Feedback and Review isCorrect: false + feedback: Feedback and Review was considered a part of effective AI policies - typeName: multipleChoice prompt: Which of the following are legitimate pieces of advice your should adhere to regarding AI policy? - shuffleOptions: true options: - answer: Always consult your legal council about the AI regulations that apply to you. isCorrect: false + feedback: - answer: It is not enough to build a policy, support systems need to be in place to support individuals in their ability to follow that policy. isCorrect: false - answer: You should seek advice from a variety of experts and sources when building a comprehensive AI policy diff --git a/coursera_quizzes/04b-AI_Policy-ai_regs_and_laws_quiz.md.yml b/coursera_quizzes/04b-AI_Policy-ai_regs_and_laws_quiz.md.yml index f7e5c3c9..6d5733f4 100644 --- a/coursera_quizzes/04b-AI_Policy-ai_regs_and_laws_quiz.md.yml +++ b/coursera_quizzes/04b-AI_Policy-ai_regs_and_laws_quiz.md.yml @@ -5,25 +5,27 @@ options: - answer: To ban all AI systems isCorrect: false + feedback: Incorrect, it does not ban all AI systems - answer: To limit the capacity of AI systems to cause harm isCorrect: true + feedback: Correct! THat is the goal of the EU's AI act - answer: To standardize AI systems across Europe isCorrect: false + feedback: It would be difficult and impractical to standardize AI systems at this point in time - answer: To tax companies using AI systems isCorrect: false + feedback: EU's AI act does not really concern itself with taxation - typeName: multipleChoice - prompt: Which of the following must developers provide for general purpose AI systems like GPT under the EU AI Act? - shuffleOptions: true + prompt: True or False - All open-source software is excluded from EU Act regulations options: - - answer: A list of potential biases - isCorrect: false - - answer: Documentation about the training data - isCorrect: false - - answer: Monthly accuracy reports - isCorrect: false - - answer: Cybersecurity audit reports + - answer: True -- open source software is always exempt isCorrect: false + feedback: There was an exception - go back and see the chapter + - answer: False -- high risk open source software applications are not exempt + isCorrect: true + feedback: Correct. High risk open source software are not exempt + - typeName: multipleChoice prompt: Which of these statements is true about the EU AI act? diff --git a/coursera_quizzes/04c-AI_Policy-how_to_navigate_quiz.md.yml b/coursera_quizzes/04c-AI_Policy-how_to_navigate_quiz.md.yml index bf9a2f15..f076306b 100644 --- a/coursera_quizzes/04c-AI_Policy-how_to_navigate_quiz.md.yml +++ b/coursera_quizzes/04c-AI_Policy-how_to_navigate_quiz.md.yml @@ -5,8 +5,10 @@ options: - answer: True isCorrect: false + feedback: This is not recommended. Many types of expertise are needed on an AI oversight team besides technical experts. - answer: False isCorrect: true + feedback: Correct. Many types of expertise are needed on an AI oversight team besides technical experts. - typeName: multipleChoice prompt: Which of the following were not individuals listed as part of an AI team? @@ -15,13 +17,19 @@ options: - answer: Legal council isCorrect: false + feedback: Legal council WERE listed as a part of an AI team - answer: Policy and governance analysts isCorrect: false + feedback: Policy and governance analysts WERE listed as a part of an AI team - answer: Data protection officers isCorrect: false + feedback: Data protection officers WERE listed as a part of an AI team - answer: Ethicists and oversight board members isCorrect: false + feedback: Ethicists and oversight board members WERE listed as a part of an AI team - answer: Marketers isCorrect: true + feedback: Correct, markerters were NOT listed as a required part of an AI team - answer: Trainers and educators isCorrect: false + feedback: Trainers and educators WERE listed as a part of an AI team diff --git a/coursera_quizzes/04d-AI_Policy-data_laws_quiz.md.yml b/coursera_quizzes/04d-AI_Policy-data_laws_quiz.md.yml index 86b64b6b..06e60926 100644 --- a/coursera_quizzes/04d-AI_Policy-data_laws_quiz.md.yml +++ b/coursera_quizzes/04d-AI_Policy-data_laws_quiz.md.yml @@ -5,12 +5,16 @@ options: - answer: Ideas for new inventions isCorrect: false + feedback: That isn't what copyright covers - answer: The exclusive right to publish or reproduce artistic material isCorrect: true + feedback: Yes! This is underneath copyright! - answer: Software code isCorrect: false + feedback: Software code is not underneath copyright - answer: Scientific discoveries isCorrect: false + feedback: Scientific discoveries are not underneath copyright - typeName: multipleChoice prompt: What type of information is covered under special data privacy protections? @@ -18,20 +22,27 @@ options: - answer: Social media posts isCorrect: false + feedback: Social media posts are not underneath data privacy protections - answer: Personal health information isCorrect: true + feedback: Yes! PHI is underneath data privacy protections! - answer: Public government reports isCorrect: false + feedback: Public government reports are not underneath data privacy protections - answer: News articles isCorrect: false + feedback: New articles are not underneath data privacy protections - typeName: multipleChoice - prompt: Who is most likely legally liable if an AI system was inherently flawed? + prompt: Who is most likely legally liable if an AI system was inherently flawed in its philosophy? shuffleOptions: true options: - answer: The end user of the AI system isCorrect: false - - answer: The developer of the AI algorithm + feedback: The end user didn't put the inherent flaw in the system + - answer: The developer of the open source training model used isCorrect: false + feedback: How a training model is used is not necessarily the original developers fault - answer: The designer or manufacturer of the system isCorrect: true + feedback: Yes. The designer or manufacterer of the AI system is likely liable for the inherent flaws. diff --git a/coursera_quizzes/04e-AI_Policy-creating_your_ai_policy_quiz.md.yml b/coursera_quizzes/04e-AI_Policy-creating_your_ai_policy_quiz.md.yml index c3792d86..94d05086 100644 --- a/coursera_quizzes/04e-AI_Policy-creating_your_ai_policy_quiz.md.yml +++ b/coursera_quizzes/04e-AI_Policy-creating_your_ai_policy_quiz.md.yml @@ -5,12 +5,16 @@ options: - answer: Guidelines on best practices only isCorrect: false + feedback: Guidelines is a start, but there are other items that also should be included - answer: Infrastructure to handle AI only isCorrect: false + feedback: Infrastructure is a start, but there are other items that also should be included - answer: Training materials only isCorrect: false + feedback: Training materials are a start, but there are other items that also should be included - answer: Guidelines, infrastructure, and training materials isCorrect: true + feedback: Yes! AI policy support systems should include all of these things - typeName: multipleChoice prompt: Who should be involved in creating a comprehensive AI policy? @@ -18,12 +22,16 @@ options: - answer: Chief Data Officer's office only isCorrect: false + feedback: This office should be involved, but others should be involved as well - answer: IT department only isCorrect: false + feedback: IT should be involved, but others should be involved as well - answer: Legal department only isCorrect: false + feedback: Legal department should be involved, but others should be involved as well - answer: Representatives from legal, IT, compliance, governance, diversity offices, etc. isCorrect: true + feedback: Yes! All of these individuals are persons who should be involved in policy making! And potentially more individuals. - typeName: multipleChoice prompt: What should training on an AI policy focus on? @@ -31,9 +39,13 @@ options: - answer: Explaining every detail of the policy isCorrect: false + feedback: It is not practical and unnecessary for every person to understand every detail of policy - answer: Outlining the key points to consider and detailing who they should consult for guidance isCorrect: true + feedback: Yes outlining what they need to know and encouraging communication with the proper channels is a great training strategy - answer: Punishments for non-compliance isCorrect: false + feedback: Punishments aren't a good idea if you'd like people to effectively be able to follow AI policy - answer: Using scare tactics about AI going wrong isCorrect: false + feedback: Scare tactics aren't probably what people need most in a training session diff --git a/quizzes/04b-AI_Policy-ai_regs_and_laws_quiz.md b/quizzes/04b-AI_Policy-ai_regs_and_laws_quiz.md index c37ae9a2..fdb80865 100644 --- a/quizzes/04b-AI_Policy-ai_regs_and_laws_quiz.md +++ b/quizzes/04b-AI_Policy-ai_regs_and_laws_quiz.md @@ -9,11 +9,9 @@ B) To limit the capacity of AI systems to cause harm c) To standardize AI systems across Europe d) To tax companies using AI systems -? Which of the following must developers provide for general purpose AI systems like GPT under the EU AI Act? -a) A list of potential biases -b) Documentation about the training data -c) Monthly accuracy reports -d) Cybersecurity audit reports +? True or False - All open-source software is excluded from EU Act regulations +a) True -- open source software is always exempt +B) False -- high risk open source software applications are not exempt ? Which of these statements is true about the EU AI act? A) There are exceptions to the facial scraping ban that allow law enforcement and intelligence agencies to use AI applications for facial recognition purposes. From c17def6d408d994927311ebd20e959748e56a9fe Mon Sep 17 00:00:00 2001 From: Candace Savonen Date: Mon, 18 Dec 2023 15:03:23 -0500 Subject: [PATCH 8/8] More feedback prompts --- .../04-AI_Policy_overall_quiz.md.yml | 22 +++++++++++++++++-- quizzes/04-AI_Policy_overall_quiz.md | 3 +-- 2 files changed, 21 insertions(+), 4 deletions(-) diff --git a/coursera_quizzes/04-AI_Policy_overall_quiz.md.yml b/coursera_quizzes/04-AI_Policy_overall_quiz.md.yml index a848f9da..e4f7543f 100644 --- a/coursera_quizzes/04-AI_Policy_overall_quiz.md.yml +++ b/coursera_quizzes/04-AI_Policy_overall_quiz.md.yml @@ -43,17 +43,22 @@ options: - answer: Always consult your legal council about the AI regulations that apply to you. isCorrect: false - feedback: + feedback: This is true, but there's more - answer: It is not enough to build a policy, support systems need to be in place to support individuals in their ability to follow that policy. isCorrect: false + feedback: This is true, but there's more - answer: You should seek advice from a variety of experts and sources when building a comprehensive AI policy isCorrect: false + feedback: This is true, but there's more - answer: A and B are true isCorrect: false + feedback: A and B are true, but there's more - answer: B and C are true isCorrect: false + feedback: B and C are true, but there's more - answer: All of the above are true isCorrect: true + feedback: Yes! All of these above are true statements! - typeName: multipleChoice prompt: What is one concern around the use of copyrighted data to train AI models? @@ -61,12 +66,16 @@ options: - answer: The models may produce biased results isCorrect: false + feedback: Bias is important to look out for but is not necessarily related to copyrighted data - answer: The models may violate intellectual property rights isCorrect: true + feedback: Correct, this has and will continue to be a concern - answer: The models may be insecure isCorrect: false + feedback: Security is important for but is not necessarily related to copyrighted data - answer: The models will use too much computer processing power isCorrect: false + feedback: Efficient computing is important but is not necessarily related to copyright data - typeName: multipleChoice prompt: What does the text recommend to allow an AI policy to stay relevant as technology changes quickly? @@ -74,12 +83,16 @@ options: - answer: Update it daily isCorrect: false + feedback: Daily would be a lot to do. This isn't what was recommended. - answer: Make it a short-term policy only isCorrect: false + feedback: This isn't what was recommended. What would happen after that policy expired? - answer: Create it as an ongoing living document isCorrect: true + feedback: Yes! Continue to modify and update the policy to better meet the changing needs and landscape of AI - answer: Don't create a policy at all isCorrect: false + feedback: This wouldn't help! Policy is needed! - typeName: multipleChoice prompt: An effective AI policy... @@ -88,11 +101,16 @@ options: - answer: Is agile and readily adjusted as the landscape of technology changes isCorrect: false + feedback: This is true, but there is more - answer: Is accompanied by training, education and infrastructure isCorrect: false + feedback: This is true, but there is more - answer: Has been created by a team of experts weighing in isCorrect: false + feedback: This is true, but there is more - answer: Discusses how bias and discrimination is being monitored isCorrect: false - - answer: All of the above should be included with an effective AI policy!x + feedback: This is true, but there is more + - answer: All of the above should be included with an effective AI policy! isCorrect: true + feedback: Correct! All of the above are items that should be included with an effective AI policy! diff --git a/quizzes/04-AI_Policy_overall_quiz.md b/quizzes/04-AI_Policy_overall_quiz.md index f7c410a6..5b9a9f55 100644 --- a/quizzes/04-AI_Policy_overall_quiz.md +++ b/quizzes/04-AI_Policy_overall_quiz.md @@ -25,7 +25,6 @@ d) A and B are true e) B and C are true F) All of the above are true - ? What is one concern around the use of copyrighted data to train AI models? a) The models may produce biased results B) The models may violate intellectual property rights @@ -44,6 +43,6 @@ a) Is agile and readily adjusted as the landscape of technology changes b) Is accompanied by training, education and infrastructure c) Has been created by a team of experts weighing in d) Discusses how bias and discrimination is being monitored -E) All of the above should be included with an effective AI policy!x +E) All of the above should be included with an effective AI policy! {/quiz}