TikTok addiction | AI voice clone

Artificial intelligence Journalism Misinformation

Oct. 21, 2024

An educator's guide to

the week in news literacy

Hi Educator,

Like the new look? We’re testing out a new design for The Sift. We hope you find this streamlined format easy to navigate and appealing to look at. For space, as we pilot this new layout, the Kickers section this week is available in the newsletter archive. Got feedback? Be sure to drop us a note here or hit reply to this email to share your thoughts.

Thank you and enjoy! 😊

— The Sift team

In this issue

TikTok addiction | AI voice clone

Daily Do-Now

Don't miss this week's classroom-ready resource.

Top picks

An illustration of an AI bot head facing a microphone while a human head talks into the same microphone.

AI-generated audio can sound like real people.

1. Can you detect an AI voice clone?

As artificial intelligence technology advances, distinguishing between authentic voice recordings and AI-generated audio has become more challenging — and experts fear audio fakes could influence voters in the 2024 presidential election. This Washington Post story includes side-by-side audio clips that sound like former President Donald Trump and Vice President Kamala Harris — but only one is real. While some experts advise listening for parts of natural speech that are difficult for AI to mimic — such as breathing, stuttering and emphasizing certain words — others admit that even they have trouble telling the difference sometimes.

Idea:

Play the audio clips in this Washington Post report for your students to see how many can confidently tell which is authentic (without just guessing). Then discuss the implications of this technology.

Discuss:

Use the “Reflect” slide in Week 6 of the Daily Do Now resource for students to think more about this issue.

“4 ways AI can be used and abused in the 2024 election, from deepfakes to foreign interference” (Barbara A. Trish, The Conversation).

⭐ NLP Resources:

Infographic: “6 things to know about AI.”
“Introduction to Algorithms” (Checkology® virtual classroom).
👉 Check out NLP’s new AI page for more teaching resources.

2. TikTok: 35 minutes to addiction

How long does it take to become addicted to TikTok? Less than 35 minutes, or about 260 of the platform’s short-form videos, according to internal TikTok documents revealed in court after 14 attorneys general recently sued the app for allegedly harming children. Investigators also found the TikTok algorithm downgrades videos from people the platform considers not attractive and boosts videos from those who fit its beauty norms. Internal documents showed the company believes about 95% of smartphone users under age 17 use TikTok and that a tool that allows parents to set time limits for their kids failed to make an impact.

Idea:

Ask students to take an informal survey of their friends. How much time in a day or a week do their friends spend on social media? (Hint: They can check their phones for this data.) How does it make them feel? Have students share and compare their findings with the class.

Discuss:

What features make social media apps like TikTok addicting? What role do algorithms play in giving you content that interests you and holds your attention? Why do social media companies want to maximize the time users spend on their platforms? How does social media addiction affect kids and teens? What could tech companies, the government, parents and kids collectively do to combat social media addiction?

“Teen Who Wrote the Book on Breaking Social Media Addiction ‘Wants to Be an Inspiration’ (Exclusive)” (Diane Herbst, People).

3. Americans’ record-low trust in news media

Only 31% of Americans say they have confidence in accurate and fair reporting by news media, according to a recent Gallup poll. It’s the lowest since Gallup began surveying Americans about trust in media in 1972 and represents a one-point drop from the previous low of 32% recorded in 2016 and 2023.

Idea:

Ask students to evaluate their news diet. What are their sources of news? Do they intentionally seek news, avoid it or do they let news find them? Have they considered how trustworthy different sources of information about current issues and events might be? What differences do they notice between standards-based sources of news and other content (such as social media posts) about news?

Discuss:

Is low public trust a direct reflection of the quality of news reporting, or are there other factors and dynamics that might cause Americans to say they distrust news media? Are other sources of information about current events more trustworthy? How is content posted by influencers and everyday people on social media different from standards-based journalism?

Note:

For information on teens’ trust in news media, check out NLP’s new report released today: “News Literacy in America: A survey of teen information attitudes and skills.” One key finding was that most teens (65%) are somewhat trusting of news media, but nearly 7 in 10 teens perceive news organizations as intentionally adding bias.

⭐ NLP Resources:

Infographic: “In brief: News media bias.”

⭐ Featured classroom resource:

RumorGuard slide deck

Educators, these classroom-ready slide decks include teaching notes. We’ve expanded these slides for students to have a comprehensive walk-through on how to debunk false rumors.

View the RumorGuard slide deck on Google Slides.

Trump’s omission from Oregon voter pamphlet misconstrued online

A social media post reads, “Oregon voter pamphlets do not include Donald Trump. He also is not listed on the Oregon State Government website under presidential candidates. What’s going on?” and features several images from the state’s voter guide. The News Literacy Project has added a label that says “MISLEADING.”

✅ YES: The Oregon Voters’ Guide for the 2024 election does not feature former President Donald Trump.

❌ NO: The decision to leave Trump off was not made by state officials.

✅ YES: It was the Trump campaign’s choice not to participate in the pamphlet.

✅ YES: Trump will appear on Oregon’s ballot

⭐ NewsLit Takeaway

Presenting seemingly suspicious content with a leading question that suggests something is amiss is a common disinformation tactic. While posts like this typically contain some accurate information, they are frequently shared in a misleading or incomplete context. In this case, the images from Oregon’s voter pamphlet are real, but it was the Trump campaign’s decision to opt out of it and not a deliberate omission by state officials.

Image of flooded Disney World castle is AI-generated

A social media post reads, “We aren’t crying, you are… okay we are too” and includes several hashtags, including #ai. The post also features a fabricated image that shows floodwaters surrounding a castle at Walt Disney World in Florida. The News Literacy Project has added the label “AI-GENERATED IMAGE.”

❌ NO: This is not an authentic photograph of Cinderella Castle surrounded by floodwaters at Disney World in October.

✅ YES: That is an AI-generated image that falsely suggests the castle was flooded following Hurricane Milton.

⭐ NewsLit Takeaway

Content created with artificial intelligence continues to become more photorealistic, which makes it difficult to spot with the naked eye. While it’s tempting to look for visual clues, the best approach is to consider the surrounding context. Who is sharing the image? What are others saying? Have there been any credible news reports to support the viral claim?

Purposefully showing children examples of misinformation, along with the fact-checking skills they need to debunk the examples, is more effective at countering misinformation than shielding them from it, according to research by psychologists at the University of California, Berkeley.

Former President Donald Trump called for CBS to lose its broadcasting license over what he claims were deceptive edits to a 60 Minutes interview with Vice President Kamala Harris. The show pushed back, saying it edited the Harris interview the way it does any other, to be “clear, accurate and on point.”

TMZ sparked outrage and criticism over journalism ethics after it published an image of the body of former One Direction singer Liam Payne, who died on Oct. 16.

Beware of publications posing as news that arrive unsolicited in the mail: partisan “pink slime” publications are targeting voters — including the conservative Catholic Tribune papers in several swing states and the left-wing Nebraska Independent.

Voting is among the most important civic actions one can take, and yet using the word “vote” in an Instagram post could cause it to be suppressed by the platform’s algorithms — part of a business strategy to limit political content, says Washington Post tech columnist Geoffrey A. Fowler.

Misinformation about the Federal Emergency Management Agency is rampant online in the wake of Hurricanes Helene and Milton, leading to threats against and harassment of relief workers in hurricane-damaged states.

To reach mainstream Western audiences, Russian disinformation agents often fabricate primary sources — such as video, audio, images and documents — and then “launder” them by convincing influencers and international news sources that they’re legitimate.

After investigating criminal networks in her country, Guatemalan journalist Michelle Mendoza faced years of harassment and threats — fueled by misinformation — before fleeing to the U.S. More than 70% of women journalists worldwide experience online violence while working, according to an International Center for Journalists and UNESCO survey.

Thanks for reading!

Your weekly issue of The Sift is created by Susan Minichiello (@susanmini), Dan Evon (@danieljevon), Peter Adams (@PeterD_Adams), Hannah Covington (@HannahCov) and Pamela Brunskill (@PamelaBrunskill). It is edited by Mary Kane (@marykkane) and Lourdes Venard (@lourdesvenard).

Sign up to receive NLP Connections (news about our work) or switch your subscription to the non-educator version of The Sift called Get Smart About News here.