Ethics of Artificial Intelligence (AI) Practice Test

These questions are part of the practice quiz. Start practicing

• Which pair correctly matches the estimated counts of jobs displaced and created by AI by 2030?
• Which claim reflects the Amodei perspective on automation timelines?
• In a tri-gram model, which words are considered to predict the next word?
• Which statement describes static embeddings?
• What does Dennett say about p-zombies in this context?
• What is the primary aim of utilitarianism?
• What does Mill say about the meaning of an opinion without debate?
• What are context-dependent embeddings?
• What does the term 'order' denote in Markov models?
• What is a shared characteristic of large language models and word embedding models?
• Which view argues that meaning is determined by use?
• What metaphor does Marmor use to illustrate a world without privacy?
• What is a potential ethical concern when AI models access external search tools?
• Which statement best describes how data colonialism relates to generative AI?
• How does rule utilitarianism define a morally right act?
• According to Aristotle, what is central to ethics?
• Which feature is listed as an indicator in this discussion that could be used to assess AI consciousness?
• What term describes AI outputs that overly agree with a user's beliefs, even if incorrect?
• What best describes weighted distributions in this context?
• In Kant's moral philosophy, an imperfect duty is best described as a duty you must only follow sometimes.
• According to Chalmers, which combination of features would lead us to consider a system conscious?
• Which statement describes the X-Worlders regarding pain?
• Which statement accurately describes long-range dependencies in language models?
• What shift in our concept of intelligence is suggested?
• Which of the following do word embeddings capture that Markov models don't?
• What is the risk of under-attribution of AI consciousness?
• In the 'X-Worlders' scenario, what evidence of pain is observed?
• What is a neural network?
• What does Mills’ utilitarianism consider as its primary focus?
• Which form of utilitarianism is described as the most demanding?
• Which expression corresponds to the 'Bigram calculation' for P(w1,w2,w3) as described?
• In the context of LLMs, what is a token?
• What do structural protections typically involve in the context of AI policy?
• How would you best define a language model?
• How is a Markov model described as related to n-grams in the provided description?
• In the trolley problem, what would a utilitarian likely do?
• Which statement about embeddings is true?
• According to the material, treating people as a means is morally okay when what condition is met?
• What is consciousness?
• According to Marmor, privacy can be protected only if which condition is met?
• According to Marmor, metadata might not violate privacy if individuals genuinely cannot be identified; what remains the concern?
• Why might embedding models place eggplant and aubergine close in vector space?
• According to act utilitarianism in the transplant case, what should the doctor do to maximize overall good?
• What mechanism is central to the architecture of transformer models?
• Which statement best captures the relationship between intelligence and consciousness?
• What assumption does an n-gram model make about word probabilities when n=1?
• What does Butlin claim about current AI systems and consciousness?
• What is a common concern when training models on internet-scale data?
• In contractualism, when is an act wrong?
• In a bigram language model, which expression represents the probability of the word following the previous word?
• Which ethical framework emphasizes respect for human autonomy as having final value?
• What does the term 'predicting the informational environment' mean in relation to privacy protection?
• What was a key limitation of early language models that focused on word prediction?
• What is Claude?
• Is absolute privacy possible under Marmor's framework given the environment is not perfectly predictable?
• What was the stated goal of Anthropic's Claude study?
• Which embedding approach is described as prediction based embeddings?
• What is Schwitzgebel's position on mimicry and consciousness?
• Compared to large language models, traditional n-gram models lack which capabilities?
• According to the Formula of Humanity, you should ...
• What is the risk of over-attribution of consciousness?
• What does a cosine similarity score of 0 indicate about two vectors?
• If AI were to become conscious, what does the text say about its moral rights?
• Which phrase describes the internal experience of p-zombies?
• What is the 'other minds' objection to the Chinese Room?
• In a bigram model, how is the next word predicted?
• Which statement illustrates a static embedding's limitation with polysemy such as 'bank'?
• What is the main idea behind count-based embeddings?
• According to Marmor, what is the underlying interest that the right to privacy protects?
• Which item is listed as missing from n-gram models to resemble large language models?
• In a corpus split between British English and American English, what could cause a fully trained embedding model to assign high cosine similarity to eggplant and aubergine despite separation of texts?
• Which describes a problem with n-grams?
• According to the privacy discussion inspired by Marmor, if aliens monitor you constantly from afar without any chance of interaction or consent, do they violate your right to privacy?
• Which definition corresponds to the term 'heist' as used in the material?
• What does RLHF stand for?
• What is a recommended strategy for technology deployment to address ethical concerns, as discussed under 'Slow Development'?
• What ethical concern is raised by LLMs producing fluent text without genuine understanding?
• A key step in the development of Claude and ChatGPT was programming in grammatical rules and information about parts of speech.
• What is the core idea behind count-based word embeddings?
• What is a key characteristic of the Transformer that enables parallel processing?
• Which expression corresponds to the 'Trigram calculation' for P(w1,w2,w3)?
• Is it important, according to Marmor's account, for people to be aware of how technology changes the flow of information?
• During training, how do language models learn?
• Early language models often struggled with what aspect?
• What constitutes negative data in this training approach?
• What does the Robot Objection propose, and what is the corresponding reply?
• What is the key reply to the Systems Objection?
• Which statement reflects a less-cautious view of AI's impact on white-collar workers?
• What risk does Lena's portrayal emphasize about immortality achieved through mind uploading?
• What is a common argument in favor of AI passing the Turing test?
• What conclusion arises from these discussions about language tasks and understanding?
• If a company's LLM generates hate speech, then someone who works at the company must believe the content of that hate speech.
• Which of the following best defines a maxim?
• What is the role of the reward model in RLHF?
• What is the core idea behind the AI/Token Tax concept?
• The privacy risk example of government reading emails best maps to which concept?
• Which statement about chain-of-thought strategies and learning from verified rewards in AI is accurate?
• What does the 'systems objection' claim?
• What is the main proponent argument against the claim that a system lacking embodiment cannot understand language?
• Which statement best describes imperfect duties?
• In the described material, the Markov model is said to produce a series of matrices detailing movement patterns between jobs in an organization. What concept is this describing?
• In Chalmers' framing, what is a philosophical zombie (p-zombie)?
• John Stuart Mill's utilitarian views about free speech imply that social media companies should never engage in content moderation. True or False?
• During training, what process is repeated to improve the model?
• Which Picasso quote is mentioned in relation to AI and creative labor?
• In RLHF, what is Data Collection used for?
• What does Kant's Formula of Humanity require you to do with others?
• What did Trystan Goetze say about AI art?
• What does the Robot Objection specifically claim about embodiment?
• Data colonialism refers to which phenomenon?
• What does 'post training' involve for large language models?
• What is a key ethical concern associated with data colonialism in AI?
• Do utilitarians and deontologists both believe that human beings have autonomy?
• Which group were early Greek teachers of philosophy, reasoning, and public speaking known as?
• What does causal responsibility refer to in ethics?
• The 'theft of labor' concern is described as
• Which sequence correctly describes the RLHF process?
• The Lena narrative highlights concerns about which concepts regarding copies of a mind?
• Is a private conversation overheard by a friend in a public park a violation of privacy, according to this account?
• Word embeddings are a way to represent words as vectors in a high-dimensional space so that words with similar meanings end up close to each other. Which option best reflects this description in a quiz format?
• By 2030, how many jobs are estimated to be replaced by AI?
• In learning the classifier for word embeddings, what is the general objective?
• What are the three questions asked about any alleged right according to Marmor?
• What is the role of negative sampling in training the classifier?
• What is a transformer architecture notable for in NLP?
• Which statement best reflects a cautious view of AI's impact on unemployment exposure?
• Lena is likely a metaphor for which concept in discussions about AI and mind uploading?
• What is the underlying interest that the right to privacy protects?
• In Kantian ethics, which principle binds us regardless of personal desires?
• How is the 'order' of a Markov model defined in this material?
• To be creative, a product must be
• Tool calls enable AI models to perform which capability?
• What is a central aim when evaluating an AI with its own decision-making power?
• According to Mills, which pleasures are considered higher?
• What is the primary goal of alignment methods such as RLHF or a constitution?
• What does tokenization refer to in natural language processing?
• In word embeddings, what is the purpose of placing words in a high-dimensional space?
• What does RLHF stand for?
• What concern about AI-generated art is raised in the material?
• How do context-dependent embeddings differ from static embeddings?
• Which question is central to the hard problem of consciousness, as discussed?
• What kind of management tool was suggested to improve Claudius's performance in tracking customer interactions?
• What did Claude perform well in the study?
• Which statement best reflects the relationship between privacy and AI surveillance?
• In the concept of weak AI attributed to Searle, computers are best seen as what?
• How is act utilitarianism defined?
• Which statement about the categorical imperative is true?
• In the described classifier setup, which data are used as positive samples?
• Outsourcing data to AI transforms business operations by which effect?
• A cosine similarity score of -1 indicates what relation between two terms?
• What is the problem with the Turing test's philosophical move as described?
• How is a co-occurrence matrix constructed?
• In skip-gram training, what is the model's objective?
• Which ethical approach emphasizes developing good character traits such as generosity and honesty?
• What does the Turing test evaluate?
• How does RLHF work?
• Which statement about deontology is true?
• In NLP, what does a cosine similarity score of 1 signify?
• Markov models primarily capture which type of information in language models?
• Which statement best captures Kant's distinction between imperfect and perfect duties?
• What does the combination of displaced and newly created jobs suggest about the labor market by 2030?
• What is a concise statement capturing the primary critique of the p-zombie position as presented?
• Which actions are described as privacy violation examples involving electronic communications?
• In this terminology, what is a type?
• In skip-gram models, what does the window size determine?
• In Kant's universal law of nature imperative, how should you act?
• Which statement captures Mill’s claim about unchallenged opinions?
• According to Dennett's take on the hard problem, which statement best summarizes his position?
• In AI ethics, privacy rights are often described as the right to control what aspect?
• Which metric is used to measure similarity between word vectors in many embeddings?
• To create a novel output, the program starts with ______.
• In the Panopticon analogy, what is the key feature of the surveillance?
• According to Chalmers, which condition would imply a system is not conscious?
• Which statement best describes imperfect duties?
• After training, what is commonly done to adapt a model to new tasks or data?
• How do AI systems typically determine what is considered good or bad during training?
• What question about consciousness is raised by comparing animals and AI?
• Which indicator does Chalmers emphasize as signaling consciousness?
• In Kantian ethics, what is the focus of ethics?
• What does The Associationist View state about word meaning?
• What is the ultimate purpose of mapping text to vectors in these models?
• What do retraining policies advocate?
• In the Chinese Room setup, what is the purpose of the two doors and the translation book?
• According to act utilitarianism, is it morally permissible to kill the healthy donor to save five?
• Which statement best captures the primary claim of the argument against AI passing the Turing test?
• Which of the following is a primary benefit of using a Customer Relationship Management (CRM) tool?
• What is the primary goal of a language model?
• What is the key difference between act utilitarianism and rule utilitarianism?
• In utilitarian ethics, which description best expresses rule utilitarianism's approach to evaluating morality?
• What does cosine similarity measure in word embedding spaces?
• Which statement best expresses a common objection to AI passing the Turing test?
• What does a tri-gram model look at to predict the next word?
• In a bigram language model, which expression represents the probability of the word 'dog' given the previous word 'the'?
• Which scenario describes the rapid automation timeline proposed by the less-cautious view?
• Which statement best captures Dennett's view of consciousness?
• According to Mill's harm principle, free speech is permissible unless it causes what?
• According to rule utilitarianism, which rule should doctors generally adopt in the transplant case to maximize the good?
• What is the basic aim of contractualist ethics?
• By 2030, how many new jobs are estimated to be created due to AI?
• In AI alignment, what is a constitution?
• Which statement best describes a Markov model?
• What technique is discussed to improve AI transparency by showing its reasoning steps?
• What is the main takeaway from the Chinese Room argument regarding AI understanding of language?
• According to Strong AI as discussed by Searle, what is claimed about computers and minds?
• Why isn't Claude an n-gram model?
• What does normative responsibility refer to in ethics?
• A hypothetical imperative is best described as a command that applies only if what?
• Which is one reason Mill gives to tolerate free speech?
• Which statement best captures a common critique of purely look-up table AI?
• Which statement about deontology is true?
• What does the distributional hypothesis state about words that appear in similar contexts?
• Mill's beliefs on hate speech are said to reflect which type of utilitarianism?
• How can this kind of interest be protected?
• What is it to violate the right to privacy?
• According to privacy discussions, metadata can be problematic because it enables what?
• Plagiarism can involve which of the following?
• Which statement best describes perfect duties?
• What is the moral law binding us regardless of desire?
• What is the main claim of the Chinese Room thought experiment?
• Plagiarism best defined?
• What does scale refer to in the context of large language models?
• What is a token?
• Do Large Language Models pass the Turing test under extended conditions with expert judges?
• What is the key feature of the Large Language Models (LLMs)?
• Butlin's concept of indicator properties is best described as: