DAM News Round-Up – 13th February 2023


A collection of DAM related articles from around the web, hand-picked by the DAM News editorial team.

Google created an AI that can generate music from text descriptions, but won’t release it

This recent article from TechCrunch investigates Google’s new MusicLM model – an AI engine that is capable of “generating high-fidelity music from text descriptions”.  In the wake of the success of prompt-based AI tools such as OpenAI’s DALL-E and ChatGPT, it was only a matter of time before such methods were applied to music generation.  Unlike the author of the article, I remain wholly unimpressed by the results, which, to my ears, resemble early Atari/Amiga style compositions with an equally limited sonic palette.  However, the output of these early AI attempts is often akin to that of a child, and recent history has shown us that given the right developmental diet, such toddlers can evolve into teenagers with significantly improved skills in a surprisingly short space of time.

5 Ways Digital Rights Management Software Provides a Return on Investment

Licensing and rights management solution provider FADEL present five ways that a DAM system can provide increased ROI: increased productivity, faster time to market, greater asset reuse, reduced legal costs, and reduced campaign costs.  To be honest, there’s not a great deal of information to extract from this article, which is essentially a segue into peddling their Rights Cloud platform.  For a more thorough investigation, DAM News has a number of articles on the topic of Digital Asset Management and Return on Investment.

AI at the heart of the customer experience in DAM, what works?

Frédéric Sanuy of French DAM consultancy Activo takes a characteristically meandering walk through the latest developments in Artificial Intelligence.  The article includes a number of quotes from guest speakers at last year’s OnDAM conference, and explores how generative platforms such as DALL-E, ChatGPT, and Midjourney are set to have an impact on DAM, PIM and CMS – not only for synthetic content creation, but also to automate the process of generating metadata and building taxonomies.

We could run out of data to train AI language programs

This recent article from Tammy Xu of MIT Technology Review takes a look at how the huge volume of data required to train AI language models is set to run dry as early as 2026.  Until now, such data has generally been classified into two groups: low quality and high quality, with an often fuzzy boundary between them. Tammy continues to explain how, in light of a pending data drought, researchers may have to reassess these definitions, along with finding new methods to reuse and extract more information from the same data.

Inclusive Metadata: Diversity, Equity & Inclusion in Digital Asset Management

Digital Asset Management software provider Tenovos explore the topic of ethics and bias, and how DEI (diversity, ethics and inclusion) should be extended to DAM systems via a series of practical methods, including responsibly crafted metadata.  The article includes four key principles that can be used to ensure your DAM system has a better chance of navigating itself and its users through an increasingly complex moral maze.  Tips include being proactive early, not reactive later, developing an attitude of bravery by not being afraid to embrace cultural differences or difficult topics, being clear about why you are making such efforts to be inclusive, and providing warnings and disclaimers for content that users may find offensive or distressing.

Share this Article:

Leave a Reply

Your email address will not be published. Required fields are marked *