tancy firmand the NationalLottery Her- itage Fund was one of my first clients.
How would you describe the range of organisations this is aimed at, and the differing range of their digital capabilities?
MP – The Briefing is aimed at her- itage decision-makers (professional or volunteer) across the UK heritage sector, which is very diverse. That would include anything ranging from museums, galleries, libraries, parks, and sites as well as communities looking after oral histories, nature or industrial heritage. Because the AI briefing is aimed to inform about AI, and is not a guide on ‘how to use AI on your collections’, it should be a useful read for any institu- tion, regardless of where they are at on their digital transformation journey.
Is it necessary for organisations that aren’t very digitally devel- oped to have an AI awareness/ strategy? (Are there many that will struggle?)
MP – AI requires machine readable content, which today means digitising content and formatting that digitised content in a way in which the AI tool
you want to use can understand. So regardless of the task you wish to per- form, this will be the first step. Now an institution does not need to have large amounts of digital collections to make use of AI. AI is integrated into a lot of tools that most people will use every day, including for work, like their phone, emails, and various pieces of software. For this reason, heritage practitioners should have a baseline understanding of AI because they are likely to have access to the technology already.
Could you give any examples of the kinds of copyright problems that heritage organisations are likely to run into with AI?
MP – Questions of copyright and AI, in particular generative AI, have received a lot of attention in the news, and across the cre- ative and heritage industries. There are two key questions of copyright which AI raises. The first is: when an AI model is given content protected by copyright in order to ‘learn’ pattern (a process which is also called ‘training’ an AI model), does this process of ‘learning’ or ‘training’ involve an infringement of copyright? Informa- tion is extracted by the AI model, but not always copied.
So, is copyright activated? This ques- tion does not have a clear answer today, and may vary from one jurisdiction to the next. Another question is: when an AI model produces content (whether it is text, music, image, a short video clip) which looks like pre-existing works covered by copyright, is it an infringe- ment? And who is responsible for the infringement? This is another ques- tion for which we have no clear legal answers today.
Is the introduction of AI more about new copyright problems, or inten- sifying old ones?
MP – AI repeats a recurring problem for copyright which is: how do we apply high-level principles of authorship, infringement and licensing based on previous technology to a new one. In this sense, AI does not raise new questions. But AI certainly challenges the way value can be extracted from copyright-protected works, repackaged and commercialised with or without the rightsholders being involved.
Could you describe the kinds of financial or reputational risks they run by using AI and how they weigh up against the advantages?
April-May 2024
INFORMATION PROFESSIONAL 33
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56
