Week of January 27, 2026
Welcome to this week’s roundup of essential developments in AI tools and technology for authors. This edition covers major AI creativity research, new platforms for scientific writing, significant copyright developments in the EU, and important infrastructure advances shaping the AI landscape.
On 20 January, Google rolled out Personal Intelligence for the Gemini app, connecting it to your Google apps including Gmail, Photos, YouTube, and Search to provide highly personalised responses. Rather than simply retrieving information from individual apps, the system reasons across multiple data sources simultaneously. Ask Gemini about holiday recommendations and it analyses your past travel photos, reviews booking confirmations in Gmail, considers YouTube videos you’ve watched about destinations, and examines your search history.[google]
Privacy has been built into the design: Personal Intelligence is off by default, you choose which apps to connect, and you can turn it off at any time. Google confirmed that Gemini doesn’t train directly on your inbox or photo library. For authors, this cross-app reasoning could prove valuable for research workflows, connecting notes, emails, and saved references into coherent research summaries.[9to5google]
Google also announced Gemini in Chrome on 28 January, introducing auto browse capabilities that let Gemini automate browser tasks on your behalf whilst keeping you in control, alongside a new side panel for multitasking and AI-powered image editing.[gemini]
On 27 January, OpenAI launched Prism, a free AI-native workspace powered by GPT-5.2 designed specifically for scientific paper writing and research collaboration. Built on the acquired Crixet LaTeX infrastructure, Prism integrates AI directly into the research workflow, allowing scientists to draft and revise papers, search for relevant literature, and use AI to create and reason over equations, citations, and figures with real-time collaboration features.[techcrunch]
For nonfiction authors, academics, and researchers within our community, this represents a significant development. Prism offers unlimited projects and collaborators for free with personal ChatGPT accounts, removing cost barriers. The platform can convert whiteboard sketches into LaTeX, auto-generate bibliographies, and provide voice-based editing. OpenAI reports that 8.4 million science queries arrive weekly on ChatGPT, with scientists messaging 3.5 times more than average users.[aibase]
Anthropic continued expanding Claude’s capabilities throughout January. The Cowork feature, initially launched on 12 January as a desktop preview bringing agentic capabilities to knowledge work, was expanded to Pro plan users on 16 January. Cowork allows Claude to read, edit, and reorganise files, handling complex workflows without requiring continuous manual input, making it particularly useful for manuscript organisation and research compilation.[releasebot]
Notably, Anthropic has retired Claude Opus 3 as of 5 January. The replacement, Claude Opus 4.5, delivers improved performance at roughly a third less cost. Authors using Claude through API integrations should ensure their configurations point to the updated model. Researchers requiring legacy access can apply through Anthropic’s External Researcher Access Program.[theagencyjournal]
Claude Code access has also been added to Team plan Standard seats, broadening availability of Claude’s developer-focused capabilities to more organisations.[releasebot]
In a landmark vote on 28 January, the European Parliament’s Legal Affairs (JURI) Committee voted 17-3 to adopt the Voss Report on generative AI and copyright. The measures require AI companies to compensate creators and maintain detailed records of every copyrighted work used in training, including comprehensive documentation of automated data collection activities. Crucially, EU copyright law would apply to all generative AI systems operating in the European market, regardless of where training occurred.[europarl]
Rapporteur Axel Voss stated that creators deserve transparency, legal certainty, and fair compensation. The committee rejected a global flat-rate licensing system, instead calling for voluntary collective licensing agreements accessible to individual creators and SMEs. They also asked the Commission to explore whether compensation requirements could apply retroactively. The report goes to a full Parliament vote in March.[creativesunite]
For authors, this is significant. If adopted, AI providers would face legal consequences for failing to meet transparency requirements, potentially constituting copyright infringement. The EU is charting a distinctly different course from the US, where fair use rulings have produced mixed results.
The extended opt-out deadline for the historic $1.5 billion Bartz v. Anthropic settlement passed on 29 January 2026. Authors who wished to exclude themselves from the settlement and pursue independent legal action needed to file by this date. The case has been reassigned to Judge Araceli Martinez-Olguin after Judge William Alsup took inactive status.[authorsalliance]
Notable developments include journalist John Carreyrou and several additional authors opting out to pursue a separate legal action against Anthropic, Google, OpenAI, Meta, xAI, and Perplexity. The claims deadline is 30 March 2026, with a court hearing on final approval scheduled for 23 April 2026. Authors who remain in the class can expect earliest payments around August 2026.[penguin]
Speaking at the World Economic Forum in Davos on 21 January, Meta CTO Andrew Bosworth confirmed that the company’s newly formed Superintelligence Labs had delivered its first high-profile AI models internally. Bosworth described the models as “very good,” though specific capabilities and release timelines were not disclosed. The lab was established following a reorganisation of Meta’s AI efforts by CEO Mark Zuckerberg, bringing in leading researchers to compete more directly with OpenAI, Google DeepMind, and Anthropic.[yahoo]
On 26 January, Microsoft announced the Maia 200, its second-generation AI chip built on TSMC’s 3nm technology. With over 100 billion transistors, the chip delivers three times the performance of Amazon’s third-generation Trainium and outperforms Google’s seventh-generation TPU. Microsoft described it as the most efficient inference system it has ever deployed, delivering 30% better performance per cost than current hardware.[techcrunch]
Whilst this may seem distant from daily authorship, infrastructure improvements like these directly influence the cost and speed of AI tools authors rely on. More efficient chips mean lower operational costs for AI providers, which ultimately translates to more affordable and responsive writing tools.[cnbc]
A landmark study published on 21 January in Scientific Reports, led by Professor Karim Jerbi of the Universite de Montreal with contributions from AI pioneer Yoshua Bengio, tested the creativity of several leading AI models against over 100,000 human participants. The results reveal that AI systems like GPT-4 now exceed average human creative performance on divergent linguistic creativity tasks, marking a significant milestone.[nature]
However, the most creative humans, particularly the top 10%, still significantly outperform all AI models, especially on richer creative work like poetry and storytelling. The researchers also found that AI creativity depends heavily on human guidance: prompts encouraging exploration of word origins and structure produced markedly higher creativity scores.[sciencedaily]
Professor Jerbi emphasised that AI has become a powerful tool in the service of human creativity rather than a replacement for it. For authors in our community, this research reinforces what many already know through practice: AI works best as a creative collaborator, amplifying your imagination and expanding possibilities rather than generating finished work independently.[singularityhub]
Google’s January updates signal a major shift in how readers may discover books. The new Gemini agentic commerce features, announced on 11 January, enable AI to assist with complete shopping tasks from discovery through to purchase. Google’s new Business Agent lets shoppers chat with brands directly on Search, turning product discovery into dialogue.[beam.ai]
For authors, this reinforces the importance of generative engine optimisation (GEO) over traditional SEO. Ensure your book metadata, descriptions, and author bio are written clearly and factually, as AI assistants will increasingly surface these when readers ask for recommendations. Building a strong presence across multiple platforms, maintaining an active email list, and ensuring your book descriptions answer the kinds of questions a reader might ask an AI assistant are becoming essential marketing strategies.
This roundup synthesises research from industry reports, tool documentation, legal proceedings, and expert analysis to provide evidence-based guidance for authors navigating AI integration. All recommendations prioritise ethical use, authentic voice preservation, and sustainable creative practice.