Content Aware media news: April 11, 2024 (2024)

As we reach another sign-up milestone with Content Aware, thanks to all already aboard or joining anew. You are in a great company with a bit of a Who's Who of media and entertainment, from Argentina to Australia and brands such as the BBC, New York Times, News Corp, Hello!, Reach PLC, Guardian, Daily Mail, Axel Springer, Spotlight Sports Group, Hearst, Conde Nast, the FT, Independent, Economist, Red Bull, Dentsu, AS Roma, AP, AFP, Post Media, CNN, Washington Post, Sky, Global, PA, Future, ESPN, Telegraph, Business Insider, Forbes, El Periodico, PRISA, Bauer, Bloomberg, Cap Gemini, Aften Posten. Politico, the Daily (That's enough, please stop! Ed.)

If there is one thing newsrooms around the world like, it's a survey. I'd guess 9 out of 10 love them.

Into the inbox slips a piece of academic research, or the percentages of people who think a thing about a thing, and out pops free headlines and content. What's not to love? You don't even have to have an opinion on the survey: the best ones talk for themselves.

Do 8 out of 10 cats really prefer Whiskas?

Is that politician toast, or flavour of the month?

And what do the public think about, hmm, AI in the newsroom? Heck, what do journalists think about AI in the newsroom? And about AI firms paying for content made by journalists?

Well, as if by magic, three surveys this week have answered those very questions. (About AI in the newsroom and paying for content I mean, I can't comment on cats or politics, surely the internet's two most mined veins.)

First up was a hefty piece of work by The Associated Press, titled Generative AI in Journalism: The Evolution of Newswork and Ethics in a Generative Information Ecosystem. The full report PDF download is here and is definitely worth your time.

This great bit of reporting quizzed 292 journos and editorial staff about how they currently use generative AI, how they want to use it, and their concerns with regard to ethics, responsible usage, and job prospects.

And just days later, the other party in the relationship between news creator and audience was surveyed - the public.

UK polling form YouGov sat down with more than 2000 people to dig deeper into what the audience in the street thinks about AI and news media, in its timely survey AI in journalism: how would public trust in the news be affected?

You can get to it all via here at the YouGov site, or benefit from the tireless Charlotte Tobitt's digging and insight to work out what it means for the industry here.

Meanwhile across the Atlantic the News Media Alliance asked 1800 members of the public similar questions about their views of AI - as well as probing views on the subject very close to the heart of the owners and creators of media: payment by AI firms for content used to train their tech.

The results are... well, hey, it's a survey, I don't need an opinion! Let's just cherry pick the juicy bits.

WHAT JOURNALISTS THINK ABOUT A.I. IN THE NEWSROOM

• 73.8% have already used Gen AI tools, and many times for work - whether their bosses know or not
• 56% think AI should be banned from creating whole articles automatically, and have major concerns about how to use it ethically - however, they also say editors and writers have the least control over ethical use of AI
• 54% are torn ("maybe") about allowing AI firms to use their work for training LLMs, with answers such as “In theory this sounds like a good idea but it’s scary to think of not having control over how content is used", and "Is it fair to let tech giants profit on the shoulders of the reporters grinding out the hard work?”. Good question - the answer is no, of course.
• 49% have seen workflows change already because of Gen AI usage
• 20% take the tack the only good AI is a dead AI, and say the only way to use it ethically is to ban it from newsrooms completely
• 7% - remarkably few given the fears of what AI might do if left unchecked - think AI will cost them their job
• Those surveyed reeled off a laundry list of good uses to pep up workflows and make their lives "less tedious". Included in the results is a novel way of weighting (for the first time I can think of, outside of a post-pressday-pint or CMS bitching session) what exactly are the most tedious parts of the job? No surprise that "Working with Computers" was deemed much less fun than "Thinking Creatively".

WHAT AUDIENCES THINK ABOUT A.I. IN THE NEWSROOM

• 94%. I'm going straight with Charlotte Tobitt's angle here which is unignorable - especially if you care about the reputation of your publication: 94% of your audience think the downsides of AI in your newsroom will outweigh the upsides
• 48% are already expressly of the opinion AI in journalism will be entirely or predominantly a bad thing
• The younger are a little more open-minded, but not on some Kool-Aid binge: 32% of 18-24s agree with the cynics above that AI will be entirely or predominantly bad
• The public is a bit more open to some of those tedious journalist jobs being AI-ified, like spelling and grammar (73%), suggesting interview questions (51%), and content summarisation (50%)
• Hardly any, only 5% of people, think it is completely acceptable for an AI to write and publish content which simply boils down a canned press release, against 24% who think it's still completely unacceptable
• All-AI content is distinctly unwanted: 64% said the would have little or no trust in all-AI created articles
• And the juicy stat for those squeezing blood out of the AI stone: 57% say they support the idea that news publishers get money for the use of their content to train AIs

All three surveys are really worth digging into, with stacks more nuance and insight than my cherry-picking. All will help you navigate the differing challenges of selling AI as a concept internally to your business, or to your audiences (or rather, not selling it to audiences).

None of the surveys blow the lid off things in a way I think will be shocking to anyone who is in the industry. In fact they largely mirror our own (much less scientific) experience at GPP, gleaned quizzing editorial and publisher teams on this very subject prior to designing and rolling out features for our whizzo AI box of tricks GAIA.

The broad experiences in those conversations are that people in the newsrooms are super interested to see how AI can make their lives better and do things a bit more swiftly, while leaders and owners are very concerned about how - like the readers - publicised AI usage can damage the brand and undermine trust.

While giving tools that lean on AI to newsrooms is one thing, giving AI content to audiences is much more risky.

Anyway, on with this week's manually-compiled edition of Content Aware.

Hooked, lined, sunk
All it took was a tweak in Google’s Terms & Conditions for stacks of your content - by which I mean your actual personal stuff - to be fair game for harvesting by Gemini. GPP's bulldog with a biro Rob Corbidge looks at the potential fallout for mere mortals spilling over from the clash of AI titans over training data.Warning: includes classic Peter Cook.
https://www.gpp.io/comment/ai-and-the-sea-of-content-are-you-the-fish-the-sea-or-the-trawler-aPWuh7I0DLB8

AI+IQ=Au
The gold rush for content to train AI is kicking down the doors of media owners harder than a hangry bull sniffing a pile of warm corn behind a flimsy fence. As others in the industry have long said, few louder than emerging heroRicky Sutton,if you have or are sat on large piles of unique content youmusthave plans in place to get rewarded for your share in the AI companies' success, be that by joining your relevant organisations, or getting your rights people or lawyers on the case. If you think Netflix spends a lot on content, it will seem like peanuts compared to what it will cost for the most successful AI to eventually earn that title - which it will do by using better content more liberally than any of its rivals.
https://www.reuters.com/technology/inside-big-techs-underground-race-buy-ai-training-data-2024-04-05/

AI bandits' Spiderman meme
While AIs have been looting and pillaging the internet for anything that turns into vectors and revenue, is there any honour amongst thieves? Of course not, a fact Google has (for once) been the one to find out to its shock, as Open AI is rumbled scraping YouTube videos.
https://rickysutton.substack.com/p/openai-resorts-to-scraping-youtube

SEO signal noise
Back in the stone age of mid-2023 and before Google's search results meltdown and the Helpful Content Update, we used to dream fondly of the notion that we could predict which signals Google used in order to assess site pages for worthiness. SEO mystic Marie Haynes disabuses us of that notion.
https://www.mariehaynes.com/the-helpful-content-system-has-changed/

”Siri, burgle those sites for me”
New Safari AI-assisted search features are not far off, according to the rumour mills. Just as you should be concerned about what Google SGE might do to traffic, as you should with iPhones which will start to negate the need to click any links. It's Ziggy from Quantum Leap without the attitude.
https://www.macrumors.com/2024/04/10/ios-18-safari-browsing-assistant-rumor/

Shutterstocked up
And Apple spends big-ish on image and video training data too.
https://www.nasdaq.com/articles/apple-strikes-deal-with-shutterstock-for-ai-image-training

What could go wrong?
What other reasons could be leading to public fear of AI in the media? How about Elon's latest pet, Grok the AI, which took little time after being pointed at Twitter to tell the world that Iran had just started a war with Israel.
https://mashable.com/article/elon-musk-x-twitter-ai-chatbot-grok-fake-news-trending-explore

US Privacy - it's the law now
Enforceable data privacy rights sit at the heart of a landmark and widely-supported US legislative Bill unveiled this week. Like much to do with GDPR compliance, if you have that covered off you are in a good starting place in complying with the mooted American Privacy Rights Act, but the two rulesets are not clones of each other, so keep an eye out on what this means to you.
https://www.commerce.senate.gov/2024/4/committee-chairs-cantwell-mcmorris-rodgers-unveil-historic-draft-comprehensive-data-privacy-legislation

Facts in fictions
One journalist start-up is trying to turn the question of misinformation and AI back on the tech itself, starting with quizzing them on sensitive matters like upcoming elections, to see what their biases and presets are. As they say, the remedy to disinformation is better facts.
https://reutersinstitute.politics.ox.ac.uk/news/julia-angwin-fears-public-sphere-about-get-worse-ai-makes-it-easier-flood-zone-misinformation

Keep an eye on the graphs
How are your traffic charts looking, now we're over a month into the latest algorithm roll-out? Check the buzz here.
https://www.seroundtable.com/more-google-core-update-ranking-volatility-37204.html