L1
Β·
Quiz
Β·
Lab
L2
Β·
Quiz
Β·
Lab
L3
Β·
Quiz
Β·
Lab
L4
Β·
Quiz
Β·
Lab
Module Test
Module 6 Β· Lesson 1

AI-Generated News & Synthetic Journalism

When algorithms write the first draft of history β€” and what gets lost in translation.
Can a machine report the truth if it has never experienced the world?

In 2023, Reuters published a detailed audit of its AI-assisted reporting pipeline. The agency had been using automated systems to draft earnings summaries, sports results, and commodity price alerts since 2018 β€” producing thousands of short articles per month that required minimal human editing. The experiment was largely invisible to readers.

What the audit revealed was not a newsroom replaced by machines, but one reshaped around them: human journalists increasingly spent their time on investigative work and contextual analysis while AI handled structured, data-rich routine tasks.

The Automation of Routine Reporting

Automated journalism is not new. The Associated Press began using Automated Insights' Wordsmith platform in 2014 to generate quarterly earnings reports. By 2016, AP was publishing over 3,700 corporate earnings stories per quarter β€” up from roughly 300 hand-written ones. The stories were indistinguishable from human-written text to most readers.

The Washington Post deployed its proprietary AI system Heliograf during the 2016 Rio Olympics to generate short result summaries, then expanded it to cover local political races, high school sports, and Congressional votes. Heliograf produced more than 850 articles in its first year.

These systems thrive in domains where information is structured and outcomes are unambiguous β€” financial data, box scores, weather readings. They struggle wherever context, source evaluation, ethical judgment, or narrative meaning-making are required.

Real Case

In 2023, CNET quietly published 77 AI-written personal finance articles before Futurism exposed the practice. Subsequent fact-checking found factual errors in a significant portion of those pieces. CNET paused the program, added AI disclosure labels, and revised its editorial process β€” becoming a reference case for the risks of unsupervised AI publication.

Generative AI Changes the Scale

Earlier automated journalism systems required structured data feeds as input. Large language models (LLMs) like GPT-4 and Claude changed this: they can synthesize unstructured text, simulate interviews, and generate novel prose from broad prompts. This dramatically lowers the barrier to producing plausible-sounding journalism at scale.

In 2023, NewsGuard identified over 400 websites publishing AI-generated content with little or no human oversight β€” producing thousands of articles daily on topics from politics to health. Many contained accurate information. Others recycled misinformation or hallucinated facts that looked credible.

The tension is structural: the same fluency that makes LLM-generated text readable makes errors harder to detect. A clumsy hallucination in a 2018 template system is obvious. A confident, well-structured hallucination from a 2024 LLM can fool editors.

Key Terms
Automated JournalismSoftware systems that convert structured data (financial results, sports scores) into readable prose without human writing input.
HallucinationWhen an AI generates confident, plausible-sounding content that is factually incorrect or entirely fabricated.
AI DisclosureEditorial practice of labeling content as AI-assisted or AI-generated so readers can calibrate their trust accordingly.
The Deeper Question

Journalism's authority rests partly on accountability β€” a named reporter who can be questioned, corrected, or sued. When AI generates content, accountability diffuses across the platform, the model developer, and the publisher. Readers navigating this landscape need new literacies, not just new labels.

The future of AI-generated news is not a binary of "replaced" or "not replaced" journalism. It is a continuum of delegation β€” and the critical question at each point is: who remains responsible for what gets published, and how would a reader know?

Lesson 1 Quiz

AI-Generated News & Synthetic Journalism
1. What did the Associated Press use the Wordsmith platform for beginning in 2014?
Correct. AP used Wordsmith to generate earnings summaries, scaling from ~300 hand-written to 3,700+ automated stories per quarter.
Not quite. Wordsmith was used for structured data tasks β€” specifically corporate earnings reports where data inputs were clean and outcomes unambiguous.
2. What major problem was discovered when Futurism investigated CNET's AI-written articles in 2023?
Correct. Fact-checking revealed factual errors in a meaningful share of CNET's 77 AI-generated personal finance articles, leading CNET to pause the program.
Not correct. The issue uncovered was factual inaccuracy β€” the articles looked plausible to readers, which is precisely what made the errors dangerous.
3. Why does AI "hallucination" pose a greater risk with modern LLMs than with earlier template-based systems?
Exactly. LLM hallucinations look credible because the surrounding text is well-structured and confident β€” they don't announce themselves the way a clumsy template error might.
Incorrect. The core danger is that LLM prose is fluent enough to hide errors in plain sight, unlike earlier systems whose failures were often visually apparent.

Lab 1 β€” Editorial Judgment Simulator

Practice deciding when AI-generated content is safe to publish and when it needs human intervention.

Your Assignment

You are an editor at a digital news outlet that uses AI to generate first drafts. Your AI assistant will show you draft articles or passages and you must decide: publish as-is, edit before publishing, or kill the piece entirely. Discuss your reasoning.

Start by asking to see your first AI-generated draft, or describe a type of story you want to evaluate. At least 3 exchanges unlock lab completion.
Editorial Desk AI
Lab 1
Welcome to the editorial desk. I have several AI-generated drafts queued for your review today β€” a corporate earnings brief, a breaking crime report, a health advice column, and a political analysis piece. Which would you like to evaluate first, or would you prefer I walk you through all four?
Module 6 Β· Lesson 2

Deepfakes, Synthetic Media & Trust Infrastructure

The war on reality β€” and the engineering projects trying to hold the line.
If seeing is no longer believing, what becomes the foundation of public trust?

Two days before Slovakia's parliamentary election, an audio recording began circulating on Facebook. In it, a voice convincingly resembling opposition leader Michal Ε imečka appeared to discuss buying votes and raising the price of beer. The recording was almost certainly AI-generated. Fact-checkers flagged it within hours β€” but Facebook's moratorium on election-related takedowns meant it stayed up through the vote. Ε imečka's party lost narrowly.

Whether the audio changed the outcome is unknowable. What is certain: a synthetic recording reached hundreds of thousands of voters at the exact moment it could do maximum damage, and existing trust infrastructure was not fast enough to stop it.

The Deepfake Landscape in 2024

The term "deepfake" originated in a 2017 Reddit community where users applied face-swapping neural networks to celebrities. By 2024 the technology had matured dramatically. Realistic video synthesis, voice cloning, and real-time face-replacement are available via open-source tools and commercial APIs accessible to anyone with a laptop.

A 2023 report by Sensity AI estimated that deepfake video content online was doubling approximately every six months. The vast majority (96%+ in earlier audits) targeted women with non-consensual synthetic pornography β€” a harm that precedes and dwarfs political uses, though political deepfakes attract the most media attention.

In January 2024, fake explicit images of Taylor Swift generated using AI image synthesis spread across X (formerly Twitter), reaching tens of millions of views before the platform restricted searches. The incident accelerated bipartisan U.S. Congressional discussion of the DEFIANCE Act, signed into law in July 2024, creating civil liability for non-consensual intimate imagery created with AI.

Real Case β€” Biden Robocall, January 2024

Before the New Hampshire primary, a robocall using a cloned version of President Biden's voice told Democratic voters: "Don't vote in Tuesday's primary." An estimated 5,000–25,000 voters received the call. Political consultant Steve Kramer later claimed responsibility, using a $1 AI voice service. The FCC subsequently banned AI voice cloning in robocalls under existing telecommunications law.

Technical Countermeasures

The response to synthetic media has generated its own technology ecosystem. Three main approaches are being deployed at scale:

Content Credentials (C2PA): The Coalition for Content Provenance and Authenticity, backed by Adobe, Microsoft, BBC, and others, developed a cryptographic standard that embeds tamper-evident metadata into images and video β€” recording who created content, when, and with what tools. As of 2024, Adobe's Firefly, Leica cameras, and several major news agencies have adopted C2PA tagging.

Detection Models: Companies like Hive Moderation, Reality Defender, and Intel's FakeCatcher system analyze pixel-level artifacts, inconsistent lighting, and physiological signals (blood flow patterns detectable in video) to flag synthetic content. Detection accuracy against top-tier synthesis models remains an arms race β€” detection models typically lag generation capabilities by six to eighteen months.

Watermarking: Google's SynthID, launched in 2023, embeds imperceptible watermarks directly into AI-generated images and audio at the pixel/waveform level, surviving compression and editing. The watermark is detectable by Google's systems but invisible to human viewers.

Key Terms
DeepfakeAI-synthesized media β€” video, audio, or image β€” that realistically depicts a real person doing or saying something they did not.
C2PACoalition for Content Provenance and Authenticity β€” an industry standard for cryptographic content credentials that record a file's origin and edit history.
SynthIDGoogle DeepMind's imperceptible watermarking system for AI-generated content, embedded at the generation stage.
Structural Challenge

Authentication standards work only when producers adopt them and platforms check them. A world where trusted outlets use C2PA while bad actors do not still requires audiences to know what the absence of credentials means β€” a media literacy challenge at least as large as the technical one.

Lesson 2 Quiz

Deepfakes, Synthetic Media & Trust Infrastructure
1. What is the C2PA standard designed to do?
Correct. C2PA is a provenance standard β€” it attaches tamper-evident credentials to content recording who made it, when, and with what tools.
Not quite. C2PA is a provenance/authentication standard using cryptographic metadata, not a detection or watermarking system per se.
2. What was the key limitation that allowed the synthetic Šimečka audio to remain on Facebook through the Slovak election?
Correct. Fact-checkers flagged it quickly β€” but Facebook's moratorium on election-related removals kept the audio up through the vote.
Incorrect. Fact-checkers did flag the audio promptly. The problem was platform policy β€” a moratorium on election-content takedowns β€” not detection failure.
3. Google's SynthID watermark is designed to survive what?
Correct. SynthID is embedded at the pixel/waveform level and is designed to persist through common modifications like compression and cropping.
Not quite. SynthID is engineered to survive file-level manipulations β€” compression, editing, and cropping β€” while remaining imperceptible to viewers.

Lab 2 β€” Synthetic Media Analyst

Examine real deepfake scenarios and evaluate which trust infrastructure tools apply.

Your Assignment

You are a trust & safety analyst at a major social platform. Scenarios involving synthetic media will be presented to you. For each, identify which countermeasures (C2PA, SynthID, detection models, legal frameworks) would be most relevant and what their limitations are.

Ask me to describe a synthetic media scenario, or tell me which type you'd like to explore: election interference, non-consensual imagery, financial fraud, or impersonation of public figures.
Trust & Safety AI
Lab 2
I'm ready to walk you through synthetic media scenarios. I can present cases involving election interference, non-consensual deepfake imagery, AI voice fraud in financial contexts, or real-time impersonation. Which scenario would you like to analyze first?
Module 6 Β· Lesson 3

AI Personalization, Filter Bubbles & the Attention Economy

How recommendation algorithms are rewriting what "the news" means β€” one user at a time.
When every person sees a different reality, what does "public opinion" mean?

In September 2021, the Wall Street Journal published the Facebook Files β€” thousands of internal documents leaked by whistleblower Frances Haugen. Among the most damaging findings: Facebook's own research had shown that its recommendation algorithm amplified divisive, angry content because such content drove higher engagement, and that the company had repeatedly shelved internal proposals to mitigate these effects when they appeared to reduce time-on-platform.

One internal slide summarized it starkly: "Our algorithms exploit the human brain's attraction to divisiveness."

How Recommendation Algorithms Shape Media Reality

Modern content recommendation β€” on YouTube, TikTok, Instagram, X, and Spotify β€” is driven by AI systems trained to maximize engagement metrics: clicks, watch time, shares, comments. Engagement correlates strongly with emotional arousal, novelty, and social validation. Content that is outrage-inducing, fear-provoking, or identity-affirming systematically outperforms nuanced, accurate content in these optimization landscapes.

A landmark 2019 Mozilla-funded study of YouTube's recommendation engine documented what researchers called "rabbit hole" pathways β€” sequences where users interested in mainstream political content were progressively recommended more extreme versions. A 2022 reanalysis by researchers at Princeton and NYU found the effect was more heterogeneous than initially claimed, but that ideological self-selection remained a significant driver of personalized news bubbles.

TikTok's algorithm, which lacks the social-graph-based filtering of Facebook (it does not primarily recommend content from your network), produces a different pattern: extreme homogenization of topic rather than viewpoint. Users who engage with a health-anxiety video are rapidly delivered a dense sequence of health-anxiety content regardless of political valence.

Real Case β€” YouTube Radicalization Research

Researcher Guillermo Chaslot, a former YouTube engineer, built tools to systematically map YouTube's recommendation paths. His 2019 research (published with the Guardian) documented that YouTube's algorithm recommended RT (Russia Today) content during the 2016 and 2018 elections at higher rates than mainstream outlets β€” a finding YouTube disputed but which prompted algorithm audits and changes to its "borderline content" policies.

The Personalization Paradox

Personalization creates genuine value: it helps users find relevant content in an overwhelming information environment. The problem is not personalization itself but the optimization target. Systems trained to maximize engagement do not naturally optimize for accuracy, civic value, or psychological wellbeing.

Research by Eytan Bakshy et al. at Facebook (published in Science, 2015) found that the newsfeed algorithm had a statistically significant but small effect on the ideological diversity of content users saw β€” suggesting the algorithm amplified, but did not create, self-selection effects. The debate over magnitude continues, but the directional effect is largely undisputed.

The emerging regulatory response includes the EU's Digital Services Act (DSA), which took effect in 2023 for very large online platforms. It requires platforms to offer users at least one recommendation feed not based on profiling, conduct annual risk assessments of recommendation systems' societal effects, and provide researchers with data access for independent audit.

Key Terms
Filter BubbleThe information environment created when recommendation algorithms systematically limit a user's exposure to challenging or diverse viewpoints based on past engagement.
Engagement OptimizationTraining AI recommendation systems to maximize user interaction metrics β€” clicks, watch time, shares β€” regardless of content quality or accuracy.
Digital Services Act (DSA)EU regulation requiring large platforms to offer non-profiling-based recommendation options and conduct societal risk assessments of their algorithmic systems.
The Shift Ahead

The next generation of AI media systems is being designed with explicit optimization targets beyond engagement β€” including accuracy scores, source diversity metrics, and emotional valence balancing. Whether these engineering fixes can overcome the economic incentive to maximize attention-hours remains the central unresolved question in platform governance.

Lesson 3 Quiz

AI Personalization, Filter Bubbles & the Attention Economy
1. What did Frances Haugen's leaked Facebook documents reveal about the platform's own research?
Correct. Internal research showed the algorithm exploited emotional divisiveness for engagement, and proposals to mitigate this were repeatedly deprioritized when they reduced time-on-platform.
Incorrect. The leaked documents showed the opposite β€” Facebook's own researchers documented the divisiveness amplification problem and the company consistently chose engagement metrics over remediation.
2. How does TikTok's filter bubble pattern differ from Facebook's according to research described in the lesson?
Correct. Without a social graph, TikTok produces extreme topic saturation β€” health anxiety, fitness, cooking β€” rather than primarily ideological filtering.
Not quite. The distinction made in the lesson is that TikTok's bubble is topical rather than primarily political β€” users get flooded with a narrow topic, regardless of political valence.
3. What does the EU's Digital Services Act require very large platforms to offer users regarding recommendations?
Correct. The DSA mandates a non-profiling-based recommendation option, plus risk assessments and researcher data access β€” not a full ban on personalization.
Incorrect. The DSA is more targeted: it requires an opt-out to a non-profiled feed, not elimination of recommendation systems entirely.

Lab 3 β€” Algorithm Audit Simulator

Map the recommendation pathways of a hypothetical platform and identify filter bubble risks.

Your Assignment

You are an independent researcher conducting a DSA-mandated algorithm audit for a social media platform. Use the AI assistant to explore how the platform's recommendation logic might create filter bubbles, what metrics it optimizes for, and what interventions you would recommend.

Start by specifying which platform type you want to audit (short-video, news feed, podcast, or a hypothetical one you design), or ask me to present the platform's algorithm parameters first.
Algorithm Audit AI
Lab 3
Welcome to the Algorithm Audit Lab. I can simulate the recommendation logic of a short-video platform, a social news feed, a podcast discovery engine, or a platform you design from scratch. Which would you like to audit, and what aspect interests you most β€” engagement optimization, filter bubble effects, or regulatory compliance?
Module 6 Β· Lesson 4

Ownership, Copyright & the Future Economics of Media

Who owns a story generated by a machine trained on a million human stories?
When AI can produce infinite content for near-zero cost, what is a journalist's work worth?

On December 27, 2023, The New York Times filed suit against OpenAI and Microsoft in the Southern District of New York β€” the most significant copyright action in the history of AI. The Times alleged that GPT-4 had been trained on millions of its copyrighted articles without license or payment, and that the resulting models could reproduce Times content verbatim and near-verbatim at scale, directly substituting for the newspaper's own products.

OpenAI's public response emphasized fair use and the transformative nature of model training. The case, still proceeding as of mid-2025, is widely expected to define the legal framework for AI training on copyrighted media for years to come.

The Training Data Dispute

The NYT lawsuit is the highest-profile of dozens of legal actions challenging AI training practices. Authors including John Grisham, George R.R. Martin, and Jodi Picoult joined a class action against OpenAI in 2023 through the Authors Guild. Getty Images sued Stability AI in both the UK and U.S. for training on 12 million licensed images. The recording industry's trade body, the RIAA, filed against AI music generation companies Suno and Udio in 2024.

The legal uncertainty is genuine: existing U.S. copyright doctrine has never addressed whether training an AI on copyrighted content constitutes infringement. The "fair use" defense turns on four factors including the transformative nature of the use and market substitution effects β€” both contested in the AI context.

In the EU, the AI Act and the Text and Data Mining exception in the Copyright Directive create a different framework: training on copyrighted works is permitted unless rights-holders have explicitly opted out. This opt-out model shifts the burden from AI companies seeking permission to creators seeking protection.

Real Case β€” News Publisher Licensing Deals

Rather than litigate, some publishers have chosen to negotiate. The Associated Press signed a licensing deal with OpenAI in 2023 covering access to its archive. Axel Springer (publisher of Politico and Business Insider) and Le Monde signed similar agreements. The terms of these deals are largely confidential, but they signal an emerging market for AI training data licensing β€” and the asymmetric power between large AI companies and individual publishers.

The Economics of Infinite Content

The deeper economic disruption is not legal β€” it is structural. If AI can produce unlimited content at marginal cost approaching zero, the advertising economics underpinning commercial journalism face existential pressure. Display advertising revenue for U.S. newspapers fell by more than 80% between 2006 and 2022; AI-generated content flooding search results and social feeds could accelerate the erosion of traffic-dependent revenue models.

A 2024 study by the Reuters Institute documented a growing divergence in newsroom AI strategies: large, well-funded outlets are investing in AI to increase output while maintaining human editorial oversight; smaller local newsrooms face the risk of being replaced wholesale by AI content farms. The consequence is a potential collapse of local news infrastructure β€” already severely weakened β€” at the exact moment communities need it most.

Some economists argue that AI will create new media business models: subscription-funded investigative journalism, certified-human content as a premium product, and AI-assisted personalization as a reader service. Others note that every previous "new model" for journalism has failed to fully replace lost advertising revenue. The question is whether the next transition will be managed with democratic intentionality or simply allowed to happen.

Key Terms
Fair UseA U.S. copyright doctrine allowing limited use of copyrighted material without permission under factors including transformativeness and market impact β€” currently contested in AI training cases.
Text and Data Mining (TDM) ExceptionEU copyright provision allowing AI training on copyrighted works unless the rights-holder has explicitly opted out, shifting the burden of protection to creators.
Training Data LicensingNegotiated agreements between AI developers and content producers granting rights to use copyrighted material in model training, emerging as an alternative to litigation.
The Question That Outlasts This Module

Every period of media disruption β€” the printing press, broadcast radio, the internet β€” ultimately reorganized who could speak and who could profit from speech. AI is the latest reorganization. The choices made now β€” in courts, in legislatures, in platform boardrooms, and by individual readers β€” will determine whether the reorganization produces a more diverse or more concentrated media landscape.

Lesson 4 Quiz

Ownership, Copyright & the Future Economics of Media
1. What is the core legal claim in The New York Times lawsuit against OpenAI and Microsoft?
Correct. The Times alleges unlicensed training on its copyrighted archive and market substitution β€” the model can reproduce its journalism, displacing its own product.
Incorrect. The core claim is about training data β€” that the Times' copyrighted articles were used to train GPT-4 without permission or payment, enabling the model to reproduce that content.
2. How does the EU's Text and Data Mining exception differ from the U.S. fair use approach to AI training?
Correct. The EU's opt-out model and the U.S. factor-based fair use analysis create different liability landscapes and different incentives for creators and AI companies.
Not correct. The EU TDM exception permits training by default with an opt-out mechanism. The U.S. has no equivalent specific rule β€” courts apply the four-factor fair use test to each situation.
3. According to the 2024 Reuters Institute research, what is the likely consequence of AI content production for local news specifically?
Correct. The Reuters Institute research documents a growing divergence β€” large outlets invest and adapt; already-weakened local newsrooms face replacement risk.
Incorrect. The research found the opposite: local newsrooms face disproportionate risk while large, well-resourced outlets are better positioned to manage the AI transition.

Lab 4 β€” Media Economics Strategy Advisor

Design a sustainable journalism business model for the AI era.

Your Assignment

You are advising a regional news organization with 15 staff journalists and declining ad revenue. Using what you know about AI's impact on media economics, copyright, and content production, develop a strategy for sustainable journalism in the AI era. The AI advisor will challenge your assumptions and help you think through tradeoffs.

Start by describing your hypothetical news organization β€” its beat, audience, current revenue mix β€” or ask me to assign you an organization to work with. At least 3 exchanges unlock lab completion.
Media Strategy AI
Lab 4
Welcome to the Media Economics Lab. I'll play the role of a strategic advisor helping you navigate the AI transition for a regional news organization. You can design your own outlet β€” beat, location, size, audience β€” or I can assign you one to work with. What's your starting point?

Module 6 β€” Final Test

The Future of Media Β· 15 questions Β· 80% to pass
1. The Associated Press expanded from ~300 to over 3,700 quarterly earnings stories by using which tool?
Correct. AP used Wordsmith beginning in 2014 to automate corporate earnings reporting at scale.
Incorrect. It was Automated Insights' Wordsmith β€” Heliograf was the Washington Post's separate system used for Olympics and election coverage.
2. What did CNET do immediately after Futurism exposed its AI-generated articles?
Correct. CNET paused the program and committed to disclosure β€” becoming a reference case for AI publishing ethics.
Not correct. CNET's response was to pause the program, add AI labeling, and revise its editorial oversight processes.
3. The Washington Post's Heliograf system was first deployed for coverage of which event?
Correct. Heliograf debuted generating short results summaries for the 2016 Rio Olympics before expanding to elections and local sports.
Incorrect. Heliograf was first deployed at the 2016 Rio Olympics, generating results updates, before expanding to election coverage and beyond.
4. In the context of AI content, "hallucination" refers to which phenomenon?
Correct. Hallucination is the AI term for producing fluent, confident output that is factually wrong β€” particularly dangerous when embedded in otherwise accurate text.
Incorrect. In AI, hallucination specifically means the model generates text that is stated with confidence but is factually false or entirely invented.
5. What is SynthID and who developed it?
Correct. SynthID is Google DeepMind's system that embeds invisible watermarks during content generation, surviving compression and editing.
Incorrect. SynthID was developed by Google DeepMind β€” it embeds imperceptible watermarks at generation time, not metadata labels or detection analysis.
6. The 2024 Biden robocall case used AI voice cloning to tell Democratic voters what?
Correct. The cloned Biden voice told Democratic primary voters to stay home β€” reaching an estimated 5,000–25,000 people in New Hampshire.
Incorrect. The cloned voice message specifically told Democrats "Don't vote in Tuesday's primary" β€” a direct voter suppression attempt using $1 AI voice cloning.
7. The C2PA standard is backed by which combination of organizations?
Correct. The Coalition for Content Provenance and Authenticity includes Adobe, Microsoft, BBC, and news organizations.
Incorrect. C2PA was formed by Adobe, Microsoft, BBC, and news agencies β€” a cross-industry coalition spanning tech companies, broadcasters, and press organizations.
8. What did Frances Haugen's leaked Facebook documents reveal about divisive content?
Correct. The Facebook Files showed internal awareness of the divisiveness amplification problem and a pattern of deprioritizing fixes when they reduced engagement.
Incorrect. The leaked documents revealed the opposite β€” Facebook's own researchers documented the problem and repeatedly saw their proposed solutions shelved.
9. Guillermo Chaslot's research into YouTube's recommendation algorithm found that the platform recommended which outlet's content at elevated rates during U.S. elections?
Correct. Chaslot documented elevated recommendation of RT during the 2016 and 2018 elections, prompting YouTube to examine its algorithm and introduce "borderline content" policies.
Incorrect. Chaslot's research specifically documented elevated recommendations for RT (Russia Today) during U.S. election periods.
10. The EU Digital Services Act (DSA) requires very large platforms to do which of the following?
Correct. The DSA's three main requirements are: an opt-out to non-profiled recommendations, annual societal risk assessments, and data access for independent researchers.
Incorrect. The DSA requires a non-profiling recommendation option, risk assessments, and research data access β€” it does not eliminate recommendation algorithms.
11. The New York Times lawsuit against OpenAI centers on the allegation that GPT-4 was trained on NYT articles and can do what?
Correct. Market substitution β€” the model reproducing Times journalism to the point of displacing it β€” is a core element of the fair use analysis in the lawsuit.
Incorrect. The Times alleges that GPT-4 can reproduce its copyrighted content verbatim, effectively substituting for the newspaper β€” which is central to the fair use market-harm analysis.
12. Under the EU's Text and Data Mining exception, AI training on copyrighted content is:
Correct. The TDM exception flips the burden β€” training is allowed unless the rights-holder affirmatively opts out, unlike the U.S. where courts evaluate each case on fair use factors.
Incorrect. The EU TDM exception creates an opt-out model β€” training is permitted unless the rights-holder has taken active steps to reserve their rights.
13. Which music industry body filed suit against AI music generation companies Suno and Udio in 2024?
Correct. The RIAA filed on behalf of major labels against Suno and Udio, alleging their models were trained on copyrighted recordings without license.
Incorrect. It was the RIAA β€” the Recording Industry Association of America β€” that filed the copyright suits against the AI music generation platforms.
14. TikTok's recommendation algorithm produces a different filter bubble pattern than Facebook primarily because:
Correct. Without a social graph as the primary filter, TikTok produces topical flooding β€” users get dense sequences of a single subject regardless of political orientation.
Incorrect. The structural difference is that TikTok doesn't primarily recommend content from your social network, which shifts its bubble pattern from ideological to topical.
15. According to the 2024 Reuters Institute research, which type of news organization is best positioned to manage the AI content transition?
Correct. The research documents a growing divergence β€” resource-rich organizations adapt while already-weakened smaller outlets face displacement risk.
Incorrect. The Reuters Institute research found that large, well-resourced outlets have the advantage β€” local and smaller newsrooms face disproportionate risk of displacement by AI content farms.