Is Google's slow AI immersion just boiling the publishing frog?
Google keeps integrating AI into everything and anything, striving to stay afloat and keep users within its boundaries. Although it seems bleak, not all is lost.
If a righteous publisher were sat in the audience at this week's Google I/O, it is quite possible that those around them might fear for their respiratory health. Such were the opportunities to signal dissent by coughing "B***S***" at the various pronouncements, you might have seen face masks being donned in the vicinity.
The long story short is that Google is going to put AI everywhere, in everything, in increments, until the user base finds itself floating on a tide of it, without realising they were going in the water, water with a colourful G stamped on every molecule.
The situation for our industry can be perfectly framed by resorting to our old, purposefully cryptic and misleading friend, Google's documentation. In the masterful work that is "Succeeding In AI Search", presumably written by someone with a doctorate in Comparative Ambiguity, we find the following sentence: "We've seen that when people click to a website from search results pages with AI Overviews, these clicks are higher quality, where users are more likely to spend more time on the site."
"Seen" - you can build a publishing business on "seen" can't you? SEO consultant Joe Youngblood summed up many of our feelings when he responded to Google's post on X linking to "Succeeding In AI Search", by simply asking "So why not give us this data so we can validate it? A simple UTM parameter would do the trick." It would, but it would reveal a horrible truth too. Once again, we're in a situation where the Tech Giant narrative does not match the publishing reality.
We would be fools to believe that, outside of a few publishers who can give Mountain View trouble on a political level due to their proximity to the ears of legislators, we are, as a mass, irrelevant to their broader strategy. That's a shame for them, as that same mass of publishers are the worker bees of content. At this rate, all they'll have left is synthetic honey.
Make no mistake, their intent is to keep more and more users within their proprietary boundaries. The user journey is now more of a user roundabout.
As Liz Reid, Google's Head of Search, put it in an interview with The Verge this week: "I think the search results page was a construct. The way we’ve all Googled for two decades was largely a response to the structure of the web itself: web pages in, web pages out. Good AI models are now able to get around that structure, and find and synthesize information from lots of sources. Now the question for Google, Reid says, is "is the information just presented to you, or is it presented to you in a way that feels as useful as you would like it to be?'"
Useful. Such a benign word. Such malign intent. As a wiser head than me pointed out, by "construct", Reid means "relic". Maybe Reid has a magic source farm, where quality content grows for free? If you're publishing unsourced content, that is to all intents and purposes your own, then surely you are a publisher?
Actual publishers aren't high on the priority list for people such as Reid who seem to regard anything that Google does not control as broken. Only, in this case, Google does control the web, so, who’s to blame here?
It's to OpenAI that the gaze of Google's senior people is largely directed, if only because their share options are probably juicier quicker. If that is understood as their lodestar, then everything else they are doing falls into place.
Jony be bad for business
You can guarantee they're in a panic this week after OpenAI recruited legendary Apple designer Jony Ive for a $6 billion signing-on fee, disguised as purchasing his business. Given that Ive has been brought in to work on as-yet undisclosed hardware products, the corridors at Mountain View will no doubt be abuzz with gossip, more share option comparisons, and searching questions of what their opinions should be in public. None of that gossip is about content, except in the abstract. Google is still chasing OpenAI like an elephant chasing a wasp. But, it is still the elephant.
Coupled to the fact that ChatGPT still seems to be outperforming Gemini in the actual user marketplace - in terms of feedback if not raw numbers - and you can understand their tunnel vision. That, and the looming prospect that their entire business might yet get spatchcocked and filleted by the courts in the near future.
If they see AI as their commercial life raft, then it's notable that following the I/O presentations this week the company's share price actually dipped. Is that even good or bad any more? It's hard to tell. For the rest of us, a panicked Google is like living under a precarious mountain crag with reports of earthquakes in the distance.
Amid all this, there are few bright spots for publishers, beyond the creeping feeling that we need a new internet shop window. We may have one, or at least, larger publishers may have one. Google Discover for desktop is rolling out, meaning that users utilising that search tab will start to see actual stories from actual publishers displayed, to click on if they so wish. It's a fresh opportunity to wow them and get them into your ecosystem rather than Google's, so make the best of it.
That such thin traffic gruel must be treated as a banquet is telling in itself, but it's something. At least they'll have some data on whether people prefer sourced content to a plausible AI mishmash.
Speed up the pace of product development and new launches with Glide Go, a pre-configured deployment of Glide CMS paired with a full-featured website hosted and managed by Glide.
Shift your focus to content and revenue while we manage the rest.
Request a demo to see Glide Go in action.
INMA World Congress of News Media & Kordiam Editorial Days: publishing galore
Members of the Glide team are at both the World Congress of News Media as well as the Kordiam Editorial Days in New York and Hamburg respectively, taking part in discussions on topics ranging from Gen AI , to subscriptions patterns, to newsroom tools turning problems into opportunities (hint: Kordiam & Glide already have you covered). If you're attending, swing by for a chat!
OpenAI's truthful fiction
Open AI's successful defence against a recent libel claim isn't the flex they think it is. After ChatGPT hallucinated criminal allegations against a public figure, their defence case basically rested on the argument "People know that AI answers are so untrustworthy that we can't possibly be held to blame!" Err... OK. Another admission for the funding meetings and shareholders to ponder we assume! Will Google also adopt this defence when its generated articles answers do the same and they are on the hook as publishers? Perhaps they need AI insurance.
Read
Lording it up again
As we've mentioned previously, the UK second chamber the House of Lords keeps bashing the government's proposed Data (Use and Access) Bill, and has again backed fresh amendments to protect the creative industry from AI thievery. The vote was even more convincing than the one last week, so the bill is once more sent back to the Commons legislative chamber for toughening up. Unexpected highlight of the process: Sir Elton John and Sir Paul McCartney getting stuck into politics. Do Lordships loom?
Read
Dear reader, here's your summer slop
A recent blunder involving a summer reading list featuring books that don't actually exist written by authors who do exist has amply illustrated the risks of using unverified GenAI content. The list, published in the Chicago Sun-Times and Philadelphia Inquirer, came from a trusted third-party source. People who read books are a particular type, and if the bond of trust is broken with them, they likely won't forget. For GenAI, read "laziness". The person responsible is unsurprisingly fighting to save their reputation in what can only be seen as a career-threatening over-trust of AI.
Read
MIT pulls controversial AI study
Red faces at MIT after it pulled a much hyped study into the alleged power of AI to help scientists. Does the paper's retraction mean that countless AI company funding pitch decks are now being edited to remove the spurious world-beating claims? Hmm... The Institute cited privacy laws for not disclosing details, but in today's world it almost feels irresponsible not to ask whether the reason of retraction was the use of AI in writing the material.
Read
Google's weighs in on using AI content
SEO sleuth Barry Schwartz today alights on some new Google advice on the use of AI-generated content on your site - and its potential downsides. Hold on a minute...
Read
... but will it follow its own advice?
Google has been presenting AI Overview rewritten articles as their own work, while adding no value, nor any original insight, nor anything new, all while pinching your clicks. Why do those terms sound familiar to your wincing ear? Maybe it's because they are all the metrics which Google uses to downgrade content it says is spammy and low value, such as when it used the policy to crash publisher business models just like they crashed affiliate sites. Anti-trust lawyers are no doubt taking note at the latest rules-for-thee-not-or-me scenario.
Read
Google I/O SEO aftermath
With the Google I/O conference behind us, Mark Williams-Cook shares his thoughts about AI-powered search, who will have the upper hand, SEO tactics and risks as well as potential legal challenges which may arise in the future.
Read
European Accessibility Act: are you ready?
The European Accessibility Act reaches the enforceability stage next month - are you prepped? It sets new standards for all sorts of digital services for those in the bloc, and comes (eventually) with teeth in the shape of fines. Read the full law here and make sure all your ducks are in a row!
Read
AI assistants that no one asked for
Even though having a personalised assistant sounds convenient, let's not forget that behind the friendly interface lies a familiar and worrying thing: user lock-in. AI sage David Buttle writes for Press Gazette on the precautions lawmakers should be taking now to prevent ushering in a new era of the same market dominance all over again.
Read
Google launches AI Watermark tool - but with limited impact
Google has unveiled a web-based portal that allows users to detect AI-generated content watermarked with its SynthID system. While helpful and definitely a step in the right direction, the system is still limited because of the absence of a unified watermarking standard across the industry.
Read