Context Window 2
This edition covers Google’s new AI video tool for Workspace customers, a US federal ruling dismissing a copyright suit against OpenAI, fresh data on AI adoption in publishing jobs, AI’s role in research integrity and creative writing, and award news for the IPG AI training programme.
Google released an AI video creation tool for businesses using selected Google Workspace subscription plans, with more general availability promised by the end of the year. It would be particularly useful for creating sales and marketing material and ancillary content around non-fiction books. Similar functionality exists through competitors like Synthesia, but the USP here is tight integration with existing data and documents in Google Workspace. Well worth exploring if your organisation is a Workspace customer.
A Federal judge in New York dismissed litigation brought by two media companies against OpenAI, finding that the plaintiffs had not been able to demonstrate that ChatGPT’s use of their content had caused them harm. This is significant because US courts use impact on the value of the copyrighted work as one of the factors to determine whether a Fair Use copyright exception applies. This may be a bellwether for wider litigation outcomes. Certainly, anyone thinking about longer term AI and IP strategy should be considering what a Fair Use scenario looks like.
53% of respondents to PW’s annual Jobs Report said their company was using AI, up from 23% the prior year. Many respondents expressed negative views about AI. The report quoted extensively from those views, without any acknowledgement of productivity or other benefits.
In my work around AI, I’ve tried to take a balanced view of the issues, so let me highlight a different point of view on impact: my friend Matt Webb wrote this week about the twentieth anniversary of launching the BBC’s first podcast, In Our Time. As a personal project, Matt built a companion website for the podcast using AI to assist with scraping and categorising two decades of structured data:
“I realised that, instead of doing the work myself, I could feed the web pages into OpenAI’s GPT-3 and ask it to extract the data for me. It took 20 minutes to write the code and cost me $30 to run it overnight (it would cost 30 cents today, just 18 months later)… That development time acceleration of 4 days down to 20 minutes… that’s equivalent to about 10 years of Moore’s Law cycles. That is, using generative AI like this is equivalent to computers getting 10 years better overnight.”
For academic publishers, Nature has a good piece on how publishers are dealing with research integrity when AI can create text, images and even data. There’s a quiz to spot the difference between real and generated images: I challenge anyone to do better than a coin toss at correctly classifying them.
On a similar note, another study showed most readers could not tell the difference between the works of ten great poets (including Chaucer, Shakespeare and Plath) and AI-generated verse.
Finally, some personal news: I’m very happy to say that the AI training I developed and delivered for the IPG won the award for Best Learning/Professional Development programme at the Association Excellence Awards last week, and has also been shortlisted for the Futurebook Excellence Award, announced on 25 November. Thanks to all of my colleagues at the IPG and the organisations I’ve delivered training and presentations to in the last year.
This was originally published in my email newsletter. To receive weekly updates on how AI is affecting the publishing industry, sign up here.