What are we reading in AI?

We may have reached peak hype for Generative AI but development shows no signs of slowing. It is crucial to hone an ability for discerning signal from noise as opportunities and threats from both demand and supply sides continue to evolve. Relevant news from last week include comprehensive overviews of generative AI from journalistic, venture capital and strategic perspectives; Big Tech product releases; copyright disputes; and cutting-edge developments in multi-modal applications and new problem domains.

• FT Strategies hosted the launch of Generating Change, JournalismAI’s latest research report which surveyed 105 news organisations from 46 countries about their engagement with AI and associated technologies. JournalismAI is a global initiative of Polis - the journalism think tank at the London School of Economics and Political Science (LSE) - and is supported by the Google News Initiative. The research showed that most news organisations see opportunities for generative AI to support their journalism, and recognise a need for ethical usage and governance (with particularly pronounced integration challenges in the Global South).
• Sequoia Capital’s overview of the generative AI ecosystem gave a holistic overview of trends, vendor markets, AI infrastructure stacks and emerging development practices, and challenges to scaling (acquisition is rapid but engagement and retention are lacking in new AI products).
• David Caswell’s article on the Reuters Institute for the Study of Journalism (RISJ) analysed trends in generative AI, articulating different strategic approaches (e.g. focusing AI on driving efficiency vs. product differentiation) and commenting pragmatically on different types of AI projects and their infrastructure and organisational requirements.
• Big Tech companies continued to unveil improvements in their AI products. Microsoft will begin to roll out Copilot in Windows 11 updates, bringing natural language interfaces into a “simple and seamless” experience that lets you chat to your documents, data and apps across Windows 11, Microsoft 365, Edge and Bing. Google released Bard Extensions, a way to conversationally interact with Google tools “like Gmail, Docs, Drive, Google Maps, YouTube, and Google Flights and hotels”. And OpenAI announced its new DALL-E 3 model for image generation, which integrates directly with ChatGPT to suggest alternative prompt options, as well as integrating speech capabilities (will this help them to retain UX/UI dominance which is arguably a large part of ChatGPT’s success?).
• Speaking of image generation, ControlNet continues to capture the imagination of social media and open-source AI users. A technique which gained prominence as a way to generate stylish QR codes is now being applied to company logos and general art pieces, with “Spiral Town” as the latest example - a clear example of the power of generative AI to manipulate multi-modal data in a way which preserves and combines semantics and styles.
• Copyright in generative AI models remains a highly-debated topic: last week The Authors Guild and 17 authors filed a class-action lawsuit against OpenAI for copyright infringement.
• Gaming company miHoYo (behind the $1B gross annual revenue game Genshin Impact) published an overview of the emerging space of AI agents - large language model-based applications which can autonomously deconstruct and execute tasks. Building upon topics in recent publications like Interactive Simulacra of Human Behaviour (Stanford Smallville) and SHOW-1 and Showrunner Agents (The Simulation). NVIDIA’s Dr Jim Fan tweeted “If Stanford Smallville can be scaled up to Genshin's or other big MMORPG's player base, that would be quite a spectacle to watch. Both humans and AI agents would have extraordinary emergent behaviors that I couldn't possibly predict.”.
• And finally, as consultants, naturally we are interested in seeing how generative AI could impact our own work! Ethan Mollick (Professor of Entrepreneurship at Wharton) produced a working paper with BCG which clearly indicated the value of generative AI for workflow augmentation: “for 18 different tasks selected to be realistic samples of the kinds of work done at an elite consulting company, consultants using ChatGPT-4 outperformed those who did not, by a lot”. As noted in FT Strategies’ September update, we see the paper’s findings as highly relevant for the future of the creative media industry.

Think about preparing your business for experimentation, or adding structure around your efforts:

• News and media organisations are experimenting with AI.
• There are various strategic models for thinking about the opportunities (and risks) from AI; most organisations are still formulating their overarching vision.

Continue to build AI literacy and an ability for your organisation to detect and understand emerging opportunities and risks:

• Big Tech continues to fight for dominance of the consumer AI chatbot space, with multi-modal and integrated applications poised as the next incoming feature set.
• Niche image generation use cases are emerging, especially where AI can present a truly novel method rather than simply an efficiency boost.
• LLMs (with or without agent-based simulation) show promise for applications to creative knowledge work.

Stay at the forefront of AI innovation and gain access to FT Strategies resources by subscribing to our monthly newsletter below.

Sam has 5 years of experience helping clients to solve strategic business challenges using data. He has helped organisations in both the public and private sectors to define strategic roadmaps and processes for using AI. He has also designed and built innovative data solutions, working with senior stakeholders as part of critical delivery-focused teams.