1. Northeastern University researchers found delivery algorithm bias in Facebook housing ads even when advertisers selected no demographic targeting. What was the mechanism?
Correct. Delivery algorithm bias: the ML model learned who was likely to click on housing ads from historical data shaped by discrimination, then optimized delivery toward those users — skewing results without any advertiser demographic input.
Incorrect. The delivery ML model learned from historically biased engagement data and reproduced that bias when optimizing who saw the ads — a discrimination mechanism entirely within the platform's infrastructure, independent of advertiser choices.
2. The ALS Ice Bucket Challenge (2014) is a documented example of which diffusion type?
Correct. Northeastern University's analysis found the Ice Bucket Challenge spread through complex contagion: most participants joined only after receiving multiple independent nominations from different network neighbors, consistent with a high-threshold adoption model.
The Ice Bucket Challenge is a canonical complex contagion case — social validation from multiple independent sources (nominations from different friends) was required before most people participated.
3. Meta's June 2023 CIB network takedown — 7,700 accounts using AI-generated profile pictures — was detected primarily through:
Correct. Behavioural coordination patterns — not content classifiers — triggered the takedown. Individual posts looked authentic; the coordination did not. This illustrates how CIB detection differs from standard content moderation.
Behavioural signals — coordinated posting times, identical sentence structures, shared technical infrastructure — not image classifiers, triggered the takedown. Individual content looked authentic; the pattern did not.
4. What did the 2023 Nature experiment on forced cross-partisan feed diversification find about attitude change?
Correct. The study showed that exposure diversity does not straightforwardly produce belief diversity — increased access to opposing content did not translate into meaningful attitude shifts within the study window.
The Nature study found increased exposure but not attitude change — and noted some participants experienced forced cross-partisan content negatively. This is an important constraint on policy proposals that assume more diverse feeds will reduce polarization.
5. Barabási and Albert's 1999 preferential attachment model explains:
Correct. Preferential attachment — new nodes connecting with probability proportional to existing degree — naturally generates scale-free, power-law distributed networks where a few hubs hold most connections.
Barabási and Albert's preferential attachment model specifically explains the emergence of power-law degree distribution: nodes that already have many connections are more likely to attract new ones.
6. Modularity Q = 0.85 in a network partition indicates:
Correct. High modularity Q (approaching 1.0) indicates that the detected community partition captures genuinely dense intra-community clustering — the communities are meaningful, not arbitrary groupings.
Modularity Q measures how much denser intra-community edges are compared to a random graph with the same degree sequence. High Q (near 1.0) means strong, real community structure — not bot counts or community numbers.
7. What did the 2018 MIT Media Lab Science study find about how false news spread on Twitter?
Correct. The study explicitly found bots were not primarily responsible — human users spread false news faster because it was more novel, implicating the platform incentive structure rather than just bad actors.
The MIT paper found human users, not bots, primarily drove the spread advantage of false news — because it was more novel, and platform incentives reward novelty-driven engagement.
8. What new obligation does the EU DSA impose regarding recommender systems that was not previously required?
Correct. The DSA's researcher access and non-personalised feed requirements extend regulatory scrutiny from content removal to amplification decisions — the first binding legal framework to do so.
The DSA requires researcher access for independent audits and a non-personalised feed option — not source-code publication or pre-approval. It's the first law to bring amplification decisions under binding regulatory scrutiny.
9. What is a "lookalike audience" in social media advertising?
Correct. Lookalike audiences use ML to find users structurally similar to a seed group — enabling advertisers to reach new users they have no prior relationship with based on behavioral profile similarity to known converters.
Incorrect. Lookalike audiences are ML-generated: the platform finds users whose behavioral signatures most closely match a seed audience, reaching previously unknown individuals based on profile similarity.
10. The Louvain community detection algorithm was developed at:
Correct. Blondel et al. developed the Louvain algorithm at Université catholique de Louvain in 2008. It is named after the institution, not a person.
The Louvain algorithm was developed by Blondel and colleagues at Université catholique de Louvain in Belgium — hence the name. It was not developed at Facebook, MIT, or Stanford.
11. How did Facebook disclose the reach of the Internet Research Agency's content in its 2017 congressional testimony?
Correct. Facebook's October 2017 congressional disclosure estimated 126 million Americans reached by IRA content — a figure that illustrated how organic engagement amplification could scale state-linked influence operations.
Facebook's testimony stated approximately 126 million Americans were reached through about 470 Pages and 80,000 posts — demonstrating how coordinated networks can exploit engagement-based amplification to achieve influence operation scale.
12. Twitter banned all political advertising in 2019. What limitation did this policy share with Facebook's opposite approach of refusing to restrict political ads?
Correct. Both policies addressed paid placements only. The deeper mechanism — algorithmic amplification of politically extreme content in organic feeds driven by engagement optimization — remained untouched by either approach.
Incorrect. Both policies' shared limitation: they regulated paid political ads while leaving untouched the organic amplification layer, where engagement-optimized algorithms distribute political content at scale with no ad policy constraints.
13. Facebook's pixel, deployed on third-party websites, sends behavioral signals back to Meta. When does this data collection occur?
Correct. The Facebook pixel fires on third-party websites regardless of whether the user is actively on Facebook, enabling cross-site behavioral tracking that feeds into ad profile building across the web.
Incorrect. The pixel collects and transmits data whenever a user visits a pixel-equipped site — Facebook login status is irrelevant. This is the mechanism enabling cross-site behavioral surveillance.
14. Facebook's Rosetta system was primarily notable for its ability to:
Correct. Rosetta used image-text recognition to extract and evaluate text inside memes — extending moderation to a format that earlier computer-vision tools treated as opaque images.
Rosetta's key innovation was reading text inside images — catching hate speech and misinformation embedded in memes that earlier vision models missed entirely.
15. Real-time bidding (RTB) auctions complete in under 100 milliseconds. What two factors determine the winning ad?
Correct. RTB winners are determined by bid price weighted by a predicted relevance score — Ad Rank (Google) or Total Value (Meta) — creating structural pressure for engagement-maximizing content.
Incorrect. RTB auctions weight bid price against an ML-predicted relevance/quality score, meaning a lower-bidding but highly relevant advertiser can outbid a higher spender with a poor match.
16. What key conclusion did the 2018 UN Fact-Finding Mission draw about Facebook in Myanmar?
Correct. The UN used the specific phrase "determining role" — Facebook acknowledged it had insufficient Burmese-language moderation capacity, creating a policy gap with lethal consequences.
The UN report concluded Facebook played a "determining role." Inadequate Burmese-language moderation capacity meant policy gaps went undetected and unaddressed until after severe harm had occurred.
17. Facebook's "Meaningful Social Interactions" algorithm change of January 2018 was intended to prioritize what type of content?
Correct. The MSI update intended to prioritize interpersonal posts over passive broadcast content. Internal documents later revealed it inadvertently amplified divisive content because controversy drove more comments.
The MSI change aimed to surface posts that generated conversations between friends. Its unintended effect — amplifying divisive content because outrage generates more comments — was documented in internal Facebook research and later in the Facebook Papers.
18. The Brexit Twitter network study by Oxford Internet Institute found:
Correct. The Oxford researchers found extreme structural separation between Leave and Remain communities — characterized as a "dual public sphere" — before the referendum was held, with almost no bridging nodes connecting the two camps.
The Brexit study found the opposite of connectivity: structural separation was so strong that content critical of each camp almost never reached the other — a textbook echo chamber formation well before voting day.
19. Meta's Oversight Board reviews individual content cases. What is its primary structural limitation?
Correct. Fewer than 50 cases reviewed by 2024 versus millions of daily decisions — the Board's direct impact is marginal; its influence is primarily normative and through policy recommendations.
The Board does issue binding case decisions. The limitation is volume: fewer than 50 cases reviewed against millions of daily decisions means negligible direct case impact.
20. TikTok's "waterfall testing" model distributes content by:
Correct. TikTok's system tests each video with a small initial audience, measures a composite engagement score, and progressively expands reach to larger cohorts — making follower count nearly irrelevant to initial distribution.
TikTok's waterfall model starts with a small test cohort and expands reach based on performance — meaning a creator with zero followers can go viral if early engagement is strong, a fundamental departure from follower-network-based distribution.