Products
Ready to get started?
No matter where you are on your CMS journey, we're here to help. Want more info or to see Glide Publishing Platform in action? We got you.
Book a demo
As content deals signed and fear of missing the boat grows, an OpenAI leak sheds some light on its negotiation tactics. PLUS: Publisher AWS event report!
What is the nature of the content deals being struck with AI businesses by publishers at the very pinnacle of the industry?
A new leak of information from within OpenAI obtained by the estimable Adweek revealed a pitch deck supposedly used by the ambitious and tempestuous tech company to persuade those with valuable content to do deals to approve its use by ChatGPT.
Leaks are currently somewhat of a theme in the tech world, and so of course OpenAI have issued what is now becoming the templated response to Adweek's enquiry into the authenticity of the documents.
"We are engaging in productive conversations and partnerships with many news publishers around the world," an OpenAI representative told the publication. "Our confidential documents are for discussion purposes only and Adweek's reporting contains a number of mischaracterizations and outdated information."
My journalistic spideysense, well honed over the years by crafted statements from those with things to obscure, leads me to assume that a) the information is accurate, and b) someone could make a miniscule text change of no substance which would technically mean they are not lying when they say it is "outdated". It's not as though it's a company that struggles with recreating versions of existing content is it?
The deck was supposedly put together for OpenAI's Preferred Publishers Program to tantalise the premium brands the company wishes to court the most assiduously, deemed to be "select, high-quality editorial partners" according to the Adweek deck.
The upsides offered to such publishers are pretty much as one would expect. Better and more prominent link treatment within ChatGPT for example, and something termed "richer brand expression", i.e. content display features such as a branded hover link, an anchored link, and a quote box with link in the content.
Oh, and money. Branding and display is all well and good, but money is what's opening doors to partnerships (or, closing them to legal threats).
To that end, OpenAI offers two types of incentive. "Guaranteed value" is the initial licensing cash a publisher can actually put in their back pocket, and "variable value", a sum influenced by how much a publisher's content enjoys "display success" dictated by user engagement according to Adweek.
It's the ongoing use of the data to train models that OpenAI is most interested in as it tries to keep its data contemporaneous, so its bot can answer queries with up-to-date information. Without that, there's a lot of tail and no teeth.
Publishers are aware that these rights deals are somewhat of a pact-with-the-devil. I compare OpenAI to Facebook 10 years ago, when the social network was trying to pose as a helpful friend of our industry rather than a disinterested wrecker. But this time, setting aside the optics, I think OpenAI genuinely need our content much more than Facebook ever did to achieve what they want.
The majority of publishers are waiting to see how the litigation route taken by the NYT and others plays out, as well as what deals are being made. For the majority of publishers, neither route is readily available.
I'm inherently suspicious of OpenAI, I must confess. It seems to be a company which sprung from broad altruistic principles and a hopeful vision of a tech future, which rapidly slid into being an entitled bully shocked at the real world and prone to snidey comments and petty posts which waft of Facebook bickering.
Generative AI is clearly an amazing and capable technology, and used in the right way it can significantly augment human creative potential.
Regarding AI-generated search results however, the user experience is still one looking for a user case. Who actually prefers this method of information retrieval? Google has already reduced the presence of its much vaunted gen-AI Overviews feature barely days after going live with it.
The future is always to play for.
Special Report: Publishers A(WS)semble!
Around 150 senior publishing industry figures from around the world gathered at the Amazon HQ in London yesterday (Wednesday 5th) for a major publishing symposium hosted by AWS.
Delegates and panellists travelled from around the UK and Europe, as well as Tokyo, New York, Chicago, Washington, Boston, Nashville, Miami, and beyond, while publishers and sector experts included Reach PLC, Süddeutsche Zeitung, Guardian Media Group, Nikkei, Springer Nature, Sub(x), ThinkAnalytics, Informa, EPAM, and more - including your favourite CMS bods from Glide.
aws publishing symposium
Key themes under discussion covered AI-created content, emerging tech including AI paywalls and audience data platforms, first-party data strategies, and real-world experience from top publishers around the world.
A series of expert Q&A panels were kicked off by a keynote presentation from McKinsey's Clemens Schwaiger on the state of publishing in Europe, and illuminated by a revealing deep-dive on A/B testing strategies and results from Nikkei's Principal Research Engineer Norihiko Sawa. Download his presentation here.
Expert panels were:
In lively debate around the rise of AI and its disruptive effect on publishing, Haman said: “We treat AI as Augmented Intelligence rather than Artificial Intelligence - it’s there to augment human creativity it’s not there to replace it. Publishing is used to being disrupted, used to working under pressure and providing new solutions using new technologies. It’s an opportunity, to get stuck in and work out how it works for us. It’s not one-size fits all.”
Read more about the event here.
More insight into the Guardian's 'volun-scriber' success can be read here.
No matter where you are on your CMS journey, we're here to help. Want more info or to see Glide Publishing Platform in action? We got you.
Book a demo
