This has been one of the most consequential weeks for authors and AI in recent memory. The London Book Fair opened its doors at Olympia London on Monday, with AI and copyright dominating the programme. A dramatic protest by nearly 10,000 authors made headlines worldwide, while behind the scenes, reports emerged that the UK government’s contentious copyright reforms will be delayed into next year.
Around 10,000 writers, including Kazuo Ishiguro, Philippa Gregory, Richard Osman, Malorie Blackman, Mick Herron, Kate Mosse and Alan Moore, have co-published Don’t Steal This Book, a volume containing nothing but their names. Organised by composer and campaigner Ed Newton-Rex, 1,000 copies of the empty book are being distributed at this week’s London Book Fair to protest the government’s proposed “commercial research exception” to copyright law, which would allow AI companies to train on copyrighted works without permission or payment. The Society of Authors CEO Anna Ganley said: “Authors call time on the theft of their books by Big Tech. ‘Don’t Steal This Book’ demonstrates what will happen if AI companies continue to steal authors’ works: writers being unable to pay their bills, leading to empty pages and the loss of jobs across the creative industries”.theguardian+3
What this means for authors: This is the most visible collective action by UK authors against AI to date. If you haven’t already made your voice heard on the consultation, the campaign website dontstealthisbook.com offers a clear way to show support. The government’s report is due by 18 March, though that deadline may now pass without firm proposals (see item 2).
The Financial Times reported on Friday 6 March that ministers have decided to shelve proposed changes to copyright law following backlash from creative industries. Consultation responses did not favour any of the government’s proposed models for AI use of copyrighted materials, and there is no expectation that an AI bill will be included in the King’s Speech due in May. The House of Lords Communications and Digital Committee published its own report the same day, warning that the UK’s £124 billion creative sector faces a “clear and present danger” from uncredited AI use of copyrighted material and calling on ministers to rule out a text and data mining exception. Separately, the Publishers Association and publishers’ trade bodies urged the government to abandon the copyright exception entirely and instead let a licensing-first approach develop.thebookseller+5
What this means for authors: This is broadly positive news for creators. The delay means the worst-case scenario, an opt-out system that places the burden on individual authors, appears to have been averted for now. However, “kicked down the road” also means continued uncertainty. Authors should continue to engage with the Society of Authors and ALCS on these issues.
Publishers’ Licensing Services (PLS), working with the Copyright Licensing Agency (CLA) and the Authors’ Licensing and Collecting Society (ALCS), unveiled a new opt-in collective licensing framework at the London Book Fair this week. The scheme would allow publishers and authors to license their content for AI training in return for payment, offering a market-based alternative to the government’s proposed copyright exception. PLS CEO Tom West said: “The pace of change is rapid, and publishers must remain active participants in shaping how their content is used. This first stage is about engagement and collaboration”.computing+2
What this means for authors: This is significant because it demonstrates that a workable licensing model exists, undermining the tech industry’s argument that licensing copyrighted material for AI is impractical. If you are a published author, keep an eye on communications from ALCS about opting in. This could eventually become a revenue stream for writers whose works are used in AI training.
On 2 March, the US Supreme Court refused to hear the appeal of Stephen Thaler, who sought copyright registration for an image created entirely by his AI system DABUS. The decision leaves in place lower court rulings that US copyright law requires human authorship, meaning works generated solely by AI cannot be copyrighted. The ruling does not, however, address cases where a human provides substantial creative input through prompting and editing.morganlewis+3
What this means for authors: For authors using AI as a writing tool, the key distinction remains the degree of human creative input. Works you substantially direct, prompt, edit and shape should still qualify for copyright protection. Purely AI-generated text with minimal human involvement likely does not. This reinforces the importance of documenting your creative process when using AI tools. Note that UK law on this point differs: the UK already has provisions for “computer-generated works,” though how these apply to generative AI remains unclear.
On 6 March, thirteen publishers, including all five of the “Big Five” (Penguin Random House, HarperCollins, Hachette, Macmillan and Simon & Schuster), filed a copyright infringement lawsuit in New York against the pirate site Anna’s Archive. The complaint alleges that Anna’s Archive has illegally copied over 63 million books and is actively selling high-speed access to its collection to AI developers, including companies in China and Russia, in exchange for cryptocurrency payments. AAP president Maria Pallante described it as “a brazen pirate operation that steals and distributes millions of literary works while outrageously offering access to AI developers in exchange for crypto payments”.publishersweekly+2
What this means for authors: This lawsuit targets a critical node in the pipeline through which pirated books reach AI training datasets. Meta was previously found to have used Anna’s Archive content to train its Llama model. If successful, the case could make it significantly harder for AI companies to source cheap, unlicensed training data, strengthening the case for legitimate licensing arrangements.
Beyond the Don’t Steal This Book protest, this week’s London Book Fair featured an extensive programme of AI-focused sessions. Highlights included “AI on Trial: Lessons from Landmark Copyright Cases” examining Bartz v. Anthropic, Kadrey v. Meta and Getty v. Stability AI on the main stage; a session on licensing content for generative AI led by PLS, ALCS and industry experts; and “How to Survive and Thrive in a Post-Search World,” exploring how AI assistants are replacing traditional search engines for book discovery. The Society of Authors ran sessions on protecting authors’ rights and sustaining a career in the age of generative AI.publishersweekly+4
What this means for authors: The London Book Fair programme reflects how thoroughly AI has permeated every aspect of publishing. The “post-search world” session is particularly relevant: as readers increasingly use AI chatbots rather than Google to find their next read, discoverability strategies will need to shift. Authors should start thinking about how their books appear in AI-generated recommendations and summaries.
Sudowrite, the AI writing tool purpose-built for fiction authors, rolled out two significant model updates in February: Google’s Gemini 3.1 Pro and Anthropic’s Claude Sonnet 4.6, available across its Write, Draft and Plugins features. The platform also integrated Claude Opus 4.6, Anthropic’s most powerful model to date, which the team noted should be “super capable at logically demanding tasks” though they remain curious about its creative writing voice. Sudowrite has retained access to the popular Claude 3.7 Sonnet, the base model for its “Excellent Prose Mode,” beyond its scheduled deprecation.[releasebot]
What this means for authors: If you use Sudowrite, you now have access to cutting-edge AI models tuned for both creative quality and speed. The addition of Gemini 3.1 Pro is noteworthy for its massive context window, which is useful for longer manuscripts and plugin development. Try the different models to see which voice suits your writing style best. Claude Sonnet 4.6 is being praised as a strong all-round performer.
The Don’t Steal This Book campaign has drawn huge attention this week. Have you signed up, and what message would you most want to send to the UK government before the 18 March deadline? Do you favour a licensing-first approach, or would you prefer a complete ban on using copyrighted works for AI training without explicit consent? Let us know in the comments.
What it is: Spoken (spoken.press) is an AI audiobook creation platform built specifically for authors, offering multi-voice narration, duet/alternating POV narration, and voice cloning from just a short audio sample.spoken+1
What’s new: Version 1.09 introduced full ePUB support for manuscript uploads, closing the gap between professional publishing prep and audiobook production. The platform’s “Pay When Perfect” pricing model means you can narrate, re-narrate and listen to your full project before spending a penny.spoken+1
Who it’s for: Indie authors and self-publishers looking to create professional-quality audiobooks without the traditional costs. Spoken integrates ElevenLabs and Hume voice technologies, offers multi-character narration, and masters output to ACX standards.[youtube]
Pricing: Pay per project or subscribe. Free to create and experiment; you only pay at the publishing stage.
Our take: The ePUB upload feature is a genuine workflow improvement for authors who already have formatted manuscripts. The ability to test voices and narration at no cost before committing makes this particularly accessible for authors on a budget. Note that ACX/Audible does not currently accept AI-narrated content, but Spotify, Kobo, Google Play and other platforms do.
#authortoolkit #aitoolsnews #aiforauthors #publishingnews