UK Government's "AI white flag", Meta's OpenAI blockade, and another Google treasure trove

This week's Content Aware brings you a new public-domain database for AI training sets, TikTok facing a major "denied" from a US Court and Apple raising alarms over an inaccurate headline.

Published: 16:04, 19 December 2024

Lurking in the murk of social publishing
Hold your breath as our divemaster Rob peers into the abyss of social media content factories and asks if publishers would be doing themselves and their readers a favour by ignoring the platform altogether.
Read

UK Government AI thinking laid bare
A new consultation document looking at the UK position on copyright and AI content scraping has just been published, seeking responses by February 25th. Do read and make your opinion clear if you can. The current thinking leans towards an opt-out for publishers, which is effectively raising the white flag to AI crawlers. Decent summary of the document here.
Read

Call for copyright protection to boost creative sector
NLA Media Access has joined a coalition of rights holders urging the government to protect copyright and support growth in the creative and tech sectors. The group stresses that fair copyright laws are key to the success of the £100 billion creative industry.
Read

Should publishers reconsider Gen AI?
INMA suggests that blocking AI bot access to content might be a strategic mistake, saying that publishers can gain by optimising content for AI crawlers and tweaking sites and schema to promote inclusion in AI-generated search results.
Read

Harvard's public-domain AI dataset
Harvard University has released a high-quality dataset of nearly 1 million public-domain books for AI start-ups to use, created through its Institutional Data Initiative and backed by, err... Microsoft and OpenAI. Including works by authors like Shakespeare and Dickens, it is intended to level the playing field in AI development. Let's see if it prevents said start-ups pinching copyrighted stuff.
Read

AI misinformation: fact or fiction?
AI-created misinformation had limited impact on election year, with most AI content used for campaigning or other legitimate purposes being limited and non-deceptive. That's not to say misinformation does not exist: the real issue is the demand for misinformation to be created in the first place, rather than what role AI plays in it.
Read

New Google search insights revealed
An exploit has been revealed which lays bare more than 2,000 properties which Google uses to classify queries and websites, such as details on consensus scoring, site quality scores, and click probability. For SEO wizards, this is like finding pages out of Gandalf's secret spell book.
Read

Meta joins chorus to block OpenAI's for-profit shift
Meta picked their side, going with Elon Musk's effort to block OpenAI's shift from being a non-profit to for-profit organisation. The Facebook parent and AI rival argues that the new business model could give non-profit investors the same benefits as for-profit investors while gaining tax advantages, a standpoint which implies there is no motive for the action other than the public good.
Read

Apple does BBC dirty
Get ready for the need to have staff on AI watch just to keep an eye on rushed-to-market services trashing your reputation: Apple Intelligence has come under fire from the BBC and others for bungling the summarisation of news, implying that the BBC is reporting fake news of high profile events.
Read

Media struggles in the digital age: quality vs quantity
Media critic Matt Pearce discusses the flawed information economy, the financial struggles of news outlets, and the importance of direct audience control for journalism's survival in a world where digital platforms such as Facebook and TikTok prioritise quantity over quality.
Read

Short of Google being broken up and OpenAI publishing a cash-for-content manifesto, we’ll be taking a break over the festive period, so it only remains for us to wish you all well and thank you for your support - see you in 2025!