Scholarly Publishing
- Is ChatGPT corrupting peer review? Telltale words hint at AI use: A study in Nature looked at common AI adjectives in peer review reports to see if it could spot an increase in the use of GenAI tools in reviewing work (spoiler: it could). (Nature, April 2024)
- Academic publishers and AI do not need to be enemies: OUP’s David Clarke posits that the apparent standoff between GenAI firms and publishers need not end in the courtrooms, but rather that both sides should remain open and curious about shared solutions that enrich scholarship while also increasing its reach. (Times Higher Education, April 2024)
Licensing
- TIME and OpenAI Announce Strategic Content Partnership: See for example this announcement of a multi-year content deal “to bring TIME's trusted journalism to OpenAI’s products, including ChatGPT.” Skeptics question the terms and enforceability of promises like the one to cite and link back to the original source on Time.com, but it’s certainly a shift in the tide. (Time, June 2024)
- All TDM & AI Rights Reserved? Fair Use & Evolving Publisher Copyright Statements: Publishers have been updating their copyright notices to account not only for text-and-data-mining (TDM, which has been around for some time) but also AI. To get a sense of what “fair use” means in the current landscape, SPARC interviewed Kyle K. Courtney, Director of Copyright and Information Policy for Harvard Library about these changes and their enforceability. (SPARC, March 2024)
- The Interplay Between Copyright Licensing and Exclusive Rights; AI Edition: “The argument that copyright and licensing somehow stifle innovation is seemingly as old as copyright itself.” Roy Kaufman of Copyright Clearance Center looks at the role of copyright in the US and beyond, including key legal activity. (The Scholarly Kitchen, May 2024)
- Dappier raises $2M in seed funding for its AI data marketplace: Licensing is quickly becoming the growth game in AI, as evidenced by the $2 million in funding raised by a startup who helps publishers license their content to AI developers. The company not only license the content but also helps to convert and prep it for optimal ingestion. (Silicon Angle, June 2024)
Security & Trust
- Microsoft Reveals AI Security Flaw That Threatens eCommerce and Financial Services: AI doomsayers got a glimpse of their feared future when Microsoft revealed the existence of a security vulnerability in AI systems that "could pose significant risks to eCommerce platforms, financial services and customer support operations across industries." The “Skeleton Key” can bypass ethical safeguards to wreak havoc on financial systems. (PYMNTS, June 2024)
- Finding GPT-4’s mistakes with GPT-4: Everyone knows that one of the biggest risks with using GenAI tools is the risk of hallucination. So who do you get to review and verify ChatGPT output? Why not ChatGPT? That’s what Open AI have done with CriticGPT, a model based on GPT-4, which reviews ChatGPT responses to help human trainers spot mistakes. “We found that when people get help from CriticGPT to review ChatGPT code they outperform those without help 60% of the time.” Models like this should help continue to improve the output and validity of GenAI tools. (OpenAI, June 2024)
AI Product Development
- Notes on how to use LLMs in your product: Product development in the AI sphere is wildly different from traditional product development, as many of us have learned. The same stages, safeguards, and market research that have gotten us here won’t work for the next phase of products. This article discusses the new mental models, workflows, and expectations that AI demands of organizations attempting to build AI products. (Irrational Exuberance, April 2024)
- The Role of Structured Content in RAG: Garbage in, garbage out. DCL takes a look at how well-structured content and data are important when it comes to training LLMs on scholarly content. The blog also shares emerging use cases for RAG solutions in the scholarly publishing industry. (DCL, May 2024)
Keeping Up
Artificial Intelligence Glossary: The Norman Nielsen Group has published a helpful and easily referenced glossary of AI terminology, from algorithms to zero-shot prompting, for any struggling to keep up with the endless stream of new terms and acronyms.
