Digital Digest #66 - Media fragmentation; anti-disinfo efforts curbed; AI and agency models; Aussies ban social media for kids; Canada boots TikTok

Hi everyone,

Happy Monday! Welcome to this week's Digital Digest. Lots to cover this week.

Following OpenAI's recent announcement of ChatGPT Search, I got my hands on it last week. My first impression: it's powerful. Having an assistant summarize dozens of sources for me was incredibly helpful when researching gear for an upcoming bike trip, or doing quick deskside research at work. Being able to click through to sources gave me additional confidence that the tool wasn't hallucinating. While it wasn't perfect - it misrepresented some information - and thus still needs a 'human in the loop', I can see myself using it more and more over a traditional search engine.

Meanwhile, in this week's edition: The decline of anti-disinformation safeguards at social media companies; generative AI's impact on agency business models; Australia moves to bar kids from social media altogether; and Canada gives TikTok the boot.

But first, how the fragmentation of media and the rise of non-traditional channels shaped last week's US election.

1. Media fragmentation and the US election

Major elections are a hotbed of marketing innovation, and a few trends typically trickle down in short order to the corporate space. In the case of this year's US election, one of the big stand-out trends is the impact of media fragments: Podcasts with massive audiences dwarfing those of news networks, TikTok and its hyper-personalized algorithm as a news channel, and the fracturing of media consumption into fragments that each place their own lens on what they see. (Wall Street Journal - gift link)

• Why it matters: Mainstream media still have massive reach and can still set the agenda, with a waterfall effect on the rest of the media -- but people are increasingly encountering that agenda within their own ecosystem that is algorithm-driven and approaches stories from its own world view (and echo chamber).

?

Also in election-related news:

• How AI shaped the 2024 election: From ad strategy to voter sentiment analysis (Digiday)
• Trump campaign spent far less on social media than in 2020 (eMarketer)

2. The decline of anti-disinformation safeguards

Despite the absolute saturation of media with stories about AI-fueled disinformation, falsehoods abound in social media channels nowadays - with the major platforms scaling down their efforts to combat them. (Washington Post - gift link, ABC)

• Why it matters: Legal threats, the politicization of counter-disinformation efforts and financial pressures have led to top platforms curbing their efforts at countering falsehoods on their platforms. Meanwhile, the real-world impacts of these lies are easy to see - look at the race riots in the UK, or the aftermath of Hurricane Helene in Florida for examples.

Also in disinformation news:

• How election deniers sank a security conference in Georgia (New York Times - gift link)
• A new risk on the horizon: organised criminals as mercenaries of disinformation (Australian Strategic Policy Institute)
• ChatGPT told 2M people to get their election news elsewhere — and rejected 250K deepfakes (TechCrunch)
• It's shockingly easy to make a Kamala Harris deepfake (TechCrunch)

3. Generative AI's impact on agency business models

AI offers the promise of significant efficiency gains for agencies (among other benefits). In a business focused on selling time, what does that mean for their business models? (Wall Street Journal - gift link)

• Why it matters: The impact of AI on business models is something I've been chewing on for a while. Media agencies already had to grapple with it in a different way amidst the shift to programmatic media; now it's the turn of other agencies. For traditional ad agencies, it's the latest in a series of hits -- firstly they were forced to deal with the shift from splashy, expensive 30-second spots to digital video (and with knocking a zero off their production budgets at the same time); now they're facing the further commoditization of production. PR agencies aren't immune either; we're likely to see both sectors shifting to other models focused on value vs hours.

Also in generative AI news:

• Mysterious AI image generator more powerful than Midjourney breaks cover (PetaPixel)
• Introducing Oasis - an AI-generated playable version of Minecraft (Oasis, Tom's Guide)
• Precise camera controls are now available for the latest version of Runway's AI video generator (Digital Camera World, Runway
• Anthropic adds PDF support to Claude 3.5 Sonnet (Anthropic)

4. Australia moves to ban social media for kids

Australia announced its plan to introduce a social media ban for children under 16. Unlike other bans we've seen tabled, this would not include an exemption for parental consent. It would also not grandfather existing under-age social media users. Platforms including X, TikTok, Instagram and Facebook would be accountable for this, and would have a year to figure out how to prevent children under the age of 16 from accessing their apps. (AP News, Digital Watch Observatory, TechCrunch)

• Why it matters: I'm torn on this one, albeit primarily around the approach rather than the intent. Broadly, in recent years there has been increasing research showing the harmful effect of social media on children and I'm all for taking measures to minimize that. At the same time I've seen with my own kids that outright prohibitions on things, rather than supervised usage, serves only to make those things more enticing (in our house, video games are the best example of this). Meanwhile, these tools are ever-more influential broader society -- so would a blanket prohibition serve to hold kids back in the long term?

Also in social media news:

• X now lets blocked users see your public posts (Gizmodo)
• Threads now has 275m monthly active users (TechCrunch)

5. Canada bans TikTok's Canadian Operations

Canada banned TikTok from operating in the country following a national security review. Bizarrely, the government's order will prohibit TikTok from operating in the country, but will allow Canadians to continue to download and use the app. TikTok has vowed to challenge the order in court. (CBC, New York Times - gift link, TikTok, Michael Geist, TechCrunch)

• Why it matters: The rationale here doesn't hang together until you put a cynical political lens on it. If TikTok is a national security concern, as is alleged, then surely the app itself should be banned, rather than a sales and marketing operation that employs mostly Canadians? Ah, but if you consider that (a) TikTok is wildly popular, (b) 40% of Canada's population uses the app, so (c) banning the app would in all likelihood be wildly unpopular at (d) a time where Canada is barreling towards an election in which the incumbent party is not forecasted to do well.

Have a great week!

Dave