WordPress civil war, Salesforce 'safe AI', and Google's new spy project

This week's Content Aware brings you a very internet feud, another "future of AI" prediction, and Google's Project Jarvis spybot privacy nightmare.

Published: 12:39, 31 October 2024

Tech whales and content minnows
A rare site visit to the home of search by people who run websites shattered by algorithm changes gave us insight into the inner workings and outward thinkings at Google. Our would-be mahout Rob paints a picture of a distracted elephant taking a moment to interview the ants it just stood on.
Read

WordPress civil war: fallout
The WordPress vs WP Engine dispute has brought levels of attention to the niche world of website hosting and management usually only found at a CMS developer's conference, but with less merch. Outside of the wrangling tech egos, the BBC peers past the hubris into the real world effects the uncertainty is having. With this conflict going on, one might ask again who do open source CMS like WordPress belong to?
Read

Lobbyists... assemble!
UK PM Keir Starmer is doing a bit of a political two-step with publishers: just days after a report seemed to suggest the new UK Government was considering weakened copyright laws to benefit AI firms, after prodding he said AI companies should pay for content. Now the milestone first Budget is out of the way, expect a pathway to emerge soon and the lobbying to ramp up. Elsewhere he spoke on other issues relevant to the industry, including clamping down on wallet-busting SLAPP lawsuits which stifle the press.
Read

Survival of the fittest
Baidu CEO Robin Li predicts only 1% of AI companies will survive, likening the current balloon of AI startups to the dot-com bubble. Baidu, aka China's Google, is investing heavily in AI, and despite the doomsayer prediction Li is confident the tech will rapidly become a default tool for society.
Read

Salesforce unveils privacy-first generative AI
Businesses are catching on at pace to the privacy and data concerns of many commercial AI systems, and some of the market is listening. Salesforce has launched a gen AI system that "forgets" processed data to enhance privacy, labelled as a breakthrough by CEO Marc Benioff. The system connects data and actions securely and applies a "trust layer", which it says makes it a step towards safer AI integrations into workflows. Adding some hands-on insight, the way we at Glide use AI in our tools for publishers is completely premised around protecting data, so we're happy to see the emergence of this theme as a priority.
Read

"Hold the door...!"
More trouble in the Open AI paradise after key safety researcher Miles Brundage announced his exit after six years, citing increased restrictions on his remit and a shift away from safety-by-default. A reminder that Open AI started life specifically as a non-profit research org specifically to research safe AI. Adding to the smoke is former colleague and fellow departee Suchir Balaji, who says the pivoted-for-profit OpenAI has violated copyright laws and now threatens the sustainability of the internet as a whole.
Read

Project Jarvis: The great Chrome uninstalling?
Google's mooted Project Jarvis, an AI agent for automating tasks in web browsers like Chrome, looks set to trigger every data safety officer on the planet with plans for a new screenshot harvesting AI service intended to help users do things like, err... buying sponsored products. Our advice? On no account install this on a work machine: this sounds like the number one new reason for staff to get marched out of buildings holding a box of belongings, for any org which has even basic privacy obligations.
Read

OFF Radio Krakow: humans out, bots in 
Taxpayer-supported Polish radio station OFF Radio Krakow sparked controversy after replacing journalists with AI-generated presenters "to attract younger listeners" - because apparently young people are too complex for other humans to work out? The move came shortly after a round of staff dismissals and intensified calls to regulate AI in media.
Read

Google $20 Decillion fine
After a four-year legal battle, Russia hit Google with a staggering fine of $20 decillion for blocking Russian media channels. For context that is $20,000,000,000,000,000,000,000,000,000,000,000, or, horrifyingly, nearly 7.5 quintillion Elon Musks. In a reply which some wags took to be a reflection of its market dominance, Google claims the fines won't significantly affect its finances.
Read

OSI defines "Open" AI, challenging tech giants like Meta
The Open Source Initiative (OSI) has released its official definition of "open" artificial intelligence, establishing new standards to reveal their training data which pay homage to the needs of content creators and challenge tech giants like Meta and other so-called open source models.
Read