Media Makers Meet/AI and the media: Q&A with Francesco Magnocavallo, Product Officer, Contents.com, Italy. HomeAbout UsAgendaSpeakersSponsorsMeetVenueBlogContact usRegistration now closed Author: Cobus HeylPosted: 1st Feb 2023 Blog navigationPrevious Blog: AI and the media: Q&A with Jeff Chang, Condé Nast TaiwanNext Blog: AI & the Media: Q&A with Dan Pacheco, Peter A. Horvitz Endowed Chair in Journalism Innovation, USA The Di5ru.pt blog In preparation for an in-depth report on AI and the media, we spoke to several experts in the field via email. One such expert is Francesco Magnocavallo, Product Officer, Contents.com. What are your current most common use cases of AI? In the Enterprise space, it’s content automation. Clients often need efficiency in content operations, where workflows can get messy. Content production and Natural Language Generation is just the most visible part of the story. For example, use cases might revolve more around content transformation than pure free text generation for this class of clients. How else will you use AI in the coming year? Content strategy and content governance. What excites you most about the longer-term future of AI, and why? In the near to mid-term: Controllable generation and retrieval augmented generation. That is first the possibility of managing parameters on NLG, be it tone of voice, point of view or any other content model aspect. Second, real-time access to the web to have some fact-checking features. A little bit further down the line, it’s programmable browser assistants that can manage complex tasks autonomously on the web for the end user. Also, in this timeframe, automation for slides and video production. What concerns you most about the longer-term future of AI, and why? The new economics of content. It might take a while to get premium content correctly estimated again. Also, AI is known for exacerbating “the middle class” with growing inequality. What are some of the best AI tools (already available for use) that you can recommend to media companies to investigate immediately? Good multilingual paraphrasers and summarisers are still hard to come by. I’d prefer not to mention the glitter apps as everything is going to be new this year anyway. Many media companies are interested in looking at how AI/machine learning can help them with the translation of their content to expand into international markets and reach new audiences. End-to-end automation is possible. Best with some sort of man in the loop approach for proofreading: this is also our trademark at Contents.com. A people-related question. Human-in-the-loop refers to a model that requires human interaction. Can you explain how this works currently, using an example from the media, and how you think this might develop in future (for example, would this lessen as AI grows smarter)? Yes the need for human-in-the-loop is going to be less and less for many use cases going from automated to autonomous. For publishing services, human-in-the-loop is often a finishing step required to provide the client with a plug-and-play product that’s brand safe. The more hard-modelled the task, the lesser the need human intervention. In one example, an e-commerce application we managed for Accenture, this amounted to 1 minute per product listing, just to ensure 100% quality output. On the other hand, one large Italian newspaper publisher prefers to show raw machine output to newsroom people and have them apply the last retouching to their titles and summaries: you don’t want to always “outsource” sophisticated editing.