The Elo Times

𝐓𝐡𝐞 7 𝐃𝐞𝐚𝐝𝐥𝐲 𝐒𝐢𝐧𝐬 𝐨𝐟 𝐌𝐞𝐝𝐢𝐜𝐚𝐥 𝐖𝐫𝐢𝐭𝐢𝐧𝐠 𝐀𝐮𝐭𝐨𝐦𝐚𝐭𝐢𝐨𝐧 👻 Want to speed up your writing process? Good idea! But try to avoid the following common mistakes. ☠ 𝐒𝐭𝐚𝐫𝐭𝐢𝐧𝐠 𝐰𝐢𝐭𝐡 𝐭𝐡𝐞 𝐬𝐨𝐥𝐮𝐭𝐢𝐨𝐧 𝐢𝐧𝐬𝐭𝐞𝐚𝐝 𝐨𝐟 𝐭𝐡𝐞 𝐩𝐫𝐨𝐛𝐥𝐞𝐦. Wondering if there's a more modern way to organize AE tables than plowing through dozens of pages with a highlighter? Randomly throwing a language model at it will just waste your time. Better start with a detailed description of your problem. This will give you a basis for communication with experts. ☠ 𝐑𝐞𝐢𝐧𝐯𝐞𝐧𝐭𝐢𝐧𝐠 𝐭𝐡𝐞 𝐰𝐡𝐞𝐞𝐥. There are a number of companies that have been dealing with topics like the automation of clinical study reports, CSRs, for years: Yseop, Narrativa Generative AI, and of course, we at Eloquest, to name just a few. Get in touch and see what's already available or who could help you with their expertise. ☠ 𝐁𝐥𝐢𝐧𝐝 𝐭𝐫𝐮𝐬𝐭 𝐢𝐧 𝐀𝐈 𝐞𝐱𝐩𝐞𝐫𝐭𝐬 𝐰𝐡𝐨 𝐡𝐚𝐯𝐞 𝐧𝐨 𝐢𝐝𝐞𝐚 𝐚𝐛𝐨𝐮𝐭 𝐲𝐨𝐮𝐫 𝐢𝐧𝐝𝐮𝐬𝐭𝐫𝐲. Unfortunately, medical writing is too complex for that. ☠ 𝐈𝐠𝐧𝐨𝐫𝐢𝐧𝐠 𝐭𝐡𝐞 𝐞𝐱𝐩𝐞𝐫𝐢𝐞𝐧𝐜𝐞 𝐨𝐟 𝐲𝐨𝐮𝐫 𝐞𝐦𝐩𝐥𝐨𝐲𝐞𝐞𝐬. This not only scares your employees but also harms software development. You can't afford developers to develop past the users. ☠ 𝐁𝐥𝐢𝐧𝐝 𝐭𝐫𝐮𝐬𝐭 𝐢𝐧 𝐟𝐫𝐞𝐞𝐥𝐲 𝐚𝐯𝐚𝐢𝐥𝐚𝐛𝐥𝐞 𝐭𝐨𝐨𝐥𝐬. It goes without saying. ☠ 𝐀𝐥𝐰𝐚𝐲𝐬 𝐚𝐬𝐬𝐮𝐦𝐢𝐧𝐠 𝐭𝐡𝐚𝐭 𝐞𝐯𝐞𝐫𝐲𝐭𝐡𝐢𝐧𝐠 𝐲𝐨𝐮 𝐧𝐞𝐞𝐝 𝐢𝐬 𝐢𝐧𝐡𝐞𝐫𝐞𝐧𝐭 𝐢𝐧 𝐲𝐨𝐮𝐫 𝐝𝐚𝐭𝐚. Sorry to break it to you, but the chances are good that your dataset is not comprehensive enough to fully represent the affected document type. Business logic and internal agreements are all too often not documented in the input data. ☠ 𝐒𝐭𝐚𝐫𝐭𝐢𝐧𝐠 𝐚 𝐫𝐞𝐯𝐨𝐥𝐮𝐭𝐢𝐨𝐧. 𝐓𝐡𝐞 𝐝𝐞𝐭𝐚𝐢𝐥𝐬 𝐜𝐚𝐧 𝐰𝐚𝐢𝐭. It may be that your automation project has the potential to revolutionize the medical writing industry in the long run. But to reap the benefits in the near future, don't lose contact to planet earth. Happy Halloween 💘 !

𝐖𝐞 𝐬𝐡𝐨𝐮𝐥𝐝 𝐡𝐚𝐯𝐞 𝐛𝐞𝐞𝐧 𝐦𝐨𝐫𝐞 𝐚𝐟𝐫𝐚𝐢𝐝 𝐨𝐟 𝐂𝐡𝐚𝐭𝐆𝐏𝐓. When we launched Eloquest in January 2023, I was often asked whether we had any concerns about starting an #NLP company right in the middle of the #ChatGPT hype. And yes, I deliberately write "ChatGPT" and not "AI" because, back then as now, the focus was never on "AI" in general, but rather on the 1-3 currently trending, freely available tools. As an NLP company, we were—and still are—naturally often placed in the #GenAI corner. My response at the time was usually something like, "Huh? Why? We're doing something completely different." That was honest, it was true, and it still is. Seven quarters later, however, I’m well aware of the challenges that have emerged for companies like Eloquest. And they’re coming from a different angle than one might expect. 𝐈𝐭’𝐬 𝐚 𝐜𝐨𝐦𝐦𝐮𝐧𝐢𝐜𝐚𝐭𝐢𝐨𝐧 𝐩𝐫𝐨𝐛𝐥𝐞𝐦. ChatGPT was marketed as a cure-all or, at the very least, celebrated in the media as such. Everything even remotely related to text could supposedly be accelerated with it. For free, or if you wanted to make a bit of an effort, with affordable subscription plans in the range of 20-30 euros. By now, most people have realized that ChatGPT is not a cure-all and that complex business cases cannot simply be solved with #LLMs alone. But I find it harder to convey in 2024 than in 2022 that there are indeed digitalization options for complex business cases, though they need to be built first. When I sit in sales meetings, I often encounter the following mindset: 𝐖𝐞 𝐰𝐚𝐧𝐭 𝐭𝐡𝐞 𝐜𝐨𝐧𝐯𝐞𝐧𝐢𝐞𝐧𝐜𝐞 𝐚𝐧𝐝 𝐧𝐨𝐧-𝐜𝐨𝐦𝐦𝐢𝐭𝐦𝐞𝐧𝐭 𝐨𝐟 𝐂𝐡𝐚𝐭𝐆𝐏𝐓 𝐚𝐧𝐝 𝐭𝐡𝐞 𝟐𝟎-𝐞𝐮𝐫𝐨-𝐚-𝐦𝐨𝐧𝐭𝐡 𝐦𝐨𝐝𝐞𝐥, 𝐛𝐮𝐭 𝐭𝐚𝐢𝐥𝐨𝐫𝐞𝐝 𝐭𝐨 𝐨𝐮𝐫 𝐩𝐫𝐨𝐟𝐞𝐬𝐬𝐢𝐨𝐧𝐚𝐥 𝐫𝐞𝐚𝐥𝐢𝐭𝐲. How is that supposed to work? Solutions that don’t exist yet need to be developed first. While I keep harping on the need for broader general education about software development and AI, I’m not really sure how we can support that anymore: Do we need figureheads like Aleph Alpha to generate interest? Should mass media take more responsibility here? 𝐖𝐡𝐚𝐭 𝐝𝐨 𝐲𝐨𝐮 𝐭𝐡𝐢𝐧𝐤?

𝐖𝐞𝐧𝐧 𝐰𝐢𝐫 𝐯𝐨𝐧 "𝐃𝐨𝐤𝐮𝐦𝐞𝐧𝐭𝐠𝐞𝐧𝐞𝐫𝐢𝐞𝐫𝐮𝐧𝐠" 𝐬𝐩𝐫𝐞𝐜𝐡𝐞𝐧, 𝐞𝐫𝐬𝐜𝐡𝐞𝐢𝐧𝐭 𝐝𝐚𝐬 𝐯𝐢𝐞𝐥𝐞𝐧 𝐚𝐮𝐟 𝐝𝐞𝐧 𝐞𝐫𝐬𝐭𝐞𝐧 𝐁𝐥𝐢𝐜𝐤 𝐚𝐛𝐬𝐭𝐫𝐚𝐤𝐭. 

𝐂𝐚𝐧 𝐰𝐞 𝐞𝐯𝐞𝐧 𝐚𝐟𝐟𝐨𝐫𝐝 𝐭𝐨 𝐝𝐫𝐢𝐯𝐞 𝐝𝐢𝐠𝐢𝐭𝐚𝐥𝐢𝐳𝐚𝐭𝐢𝐨𝐧 𝐩𝐫𝐨𝐣𝐞𝐜𝐭𝐬 𝐰𝐢𝐭𝐡𝐨𝐮𝐭 𝐚𝐧 𝐀𝐈 𝐥𝐚𝐛𝐞𝐥?

„𝐔𝐧𝐝 𝐰𝐚𝐬 𝐠𝐞𝐧𝐚𝐮 𝐬𝐢𝐧𝐝 𝐝𝐚𝐬 𝐝𝐚𝐧𝐧 𝐟ü𝐫 𝐃𝐨𝐤𝐮𝐦𝐞𝐧𝐭𝐞, 𝐝𝐞𝐫𝐞𝐧 𝐄𝐫𝐬𝐭𝐞𝐥𝐥𝐮𝐧𝐠 𝐢𝐡𝐫 𝐚𝐮𝐭𝐨𝐦𝐚𝐭𝐢𝐬𝐢𝐞𝐫𝐭?“ Ist eine Frage, die ich oft nach Vorträgen oder in Gesprächen mit Bekannten gestellt bekomme. Eloquest ist ein NLP-Dienstleister, der sich auf die Automatisierung von Dokumentgenerierung spezialisiert hat. An sich können wir für alle erdenklich möglichen Dokumente eine Software bauen, indem wir 🚀 𝐌𝐚𝐬𝐜𝐡𝐢𝐧𝐞𝐥𝐥𝐞𝐬 𝐋𝐞𝐫𝐧𝐞𝐧, 𝐃𝐚𝐭𝐚 𝐀𝐧𝐚𝐥𝐲𝐭𝐢𝐜𝐬 𝐮𝐧𝐝 𝐋𝐋𝐌𝐬 🚀 𝐄𝐱𝐩𝐞𝐫𝐭𝐞𝐧𝐰𝐢𝐬𝐬𝐞𝐧 𝐝𝐞𝐫 𝐣𝐞𝐰𝐞𝐢𝐥𝐢𝐠𝐞𝐧 𝐁𝐫𝐚𝐧𝐜𝐡𝐞 🚀 𝐮𝐧𝐝 𝐍𝐮𝐭𝐳𝐞𝐫𝐢𝐧𝐭𝐞𝐫𝐚𝐤𝐭𝐢𝐨𝐧 𝐦𝐢𝐭𝐞𝐢𝐧𝐚𝐧𝐝𝐞𝐫 𝐤𝐨𝐦𝐛𝐢𝐧𝐢𝐞𝐫𝐞𝐧. Die Frage, die sich eher stellt, ist „𝐖𝐚𝐧𝐧 𝐥𝐨𝐡𝐧𝐭 𝐞𝐬 𝐬𝐢𝐜𝐡 𝐟ü𝐫 𝐦𝐢𝐜𝐡, 𝐞𝐢𝐧𝐞𝐧 𝐃𝐨𝐤𝐮𝐦𝐞𝐧𝐭𝐞𝐧𝐭𝐲𝐩𝐞𝐧 𝐳𝐮 𝐚𝐮𝐭𝐨𝐦𝐚𝐭𝐢𝐬𝐢𝐞𝐫𝐞𝐧?“ Es lohnt sich immer dann, wenn bei der manuellen Erstellung der Dokumente 💲 𝐫𝐞𝐝𝐮𝐧𝐝𝐚𝐧𝐭𝐞 𝐀𝐫𝐛𝐞𝐢𝐭𝐬𝐬𝐜𝐡𝐫𝐢𝐭𝐭𝐞 𝐚𝐧𝐟𝐚𝐥𝐥𝐞𝐧, 𝐝𝐢𝐞 💲 𝐳𝐞𝐢𝐭𝐢𝐧𝐭𝐞𝐧𝐬𝐢𝐯 𝐮𝐧𝐝 𝐝𝐚𝐦𝐢𝐭 💲 𝐤𝐨𝐬𝐭𝐞𝐧𝐢𝐧𝐭𝐞𝐧𝐬𝐢𝐯 𝐬𝐢𝐧𝐝. Praktischerweise sind das auch genau die Arbeitsschritte, die den meisten Mitarbeitern am wenigsten Spaß machen. Aktuell sehe ich auf Social Media oft den Slogan „Ich möchte nicht, dass KI die kreative Arbeit erledigt. Ich möchte, dass sie meine Wäsche macht.”. Genau darum geht es in Automatisierungsprojekten: 𝐋𝐚𝐬𝐬𝐭 𝐊𝐈 𝐝𝐢𝐞 𝐅𝐥𝐢𝐞ß𝐛𝐚𝐧𝐝𝐚𝐫𝐛𝐞𝐢𝐭 𝐦𝐚𝐜𝐡𝐞𝐧 ❤️.

𝐄𝐥𝐨𝐪𝐮𝐞𝐬𝐭 𝐢𝐬 𝐚𝐧 𝐍𝐋𝐏 𝐬𝐞𝐫𝐯𝐢𝐜𝐞 𝐩𝐫𝐨𝐯𝐢𝐝𝐞𝐫.  𝐍𝐋𝐏, 𝐢𝐧 𝐭𝐡𝐞 𝐜𝐨𝐧𝐭𝐞𝐱𝐭 𝐡𝐞𝐫𝐞, 𝐬𝐭𝐚𝐧𝐝𝐬 𝐟𝐨𝐫 "𝐍𝐚𝐭𝐮𝐫𝐚𝐥 𝐋𝐚𝐧𝐠𝐮𝐚𝐠𝐞 𝐏𝐫𝐨𝐜𝐞𝐬𝐬𝐢𝐧𝐠". 😣 "NLP is also viewed critically by many." 😣 "I don't think much of NLP." 😣 "NLP undermines trust." Huh? Usually, it takes a moment for me to realize that my counterpart and I are completely talking past each other, and then I get annoyed with myself because I've formulated too much from my own bubble. 👨🎓 𝐍𝐋𝐏, 𝐢𝐧 𝐭𝐡𝐞 𝐟𝐢𝐞𝐥𝐝 𝐨𝐟 𝐬𝐨𝐟𝐭𝐰𝐚𝐫𝐞 𝐝𝐞𝐯𝐞𝐥𝐨𝐩𝐦𝐞𝐧𝐭, 𝐬𝐭𝐚𝐧𝐝𝐬 𝐟𝐨𝐫 𝐍𝐚𝐭𝐮𝐫𝐚𝐥 𝐋𝐚𝐧𝐠𝐮𝐚𝐠𝐞 𝐏𝐫𝐨𝐜𝐞𝐬𝐬𝐢𝐧𝐠. Natural Language Processing is a branch of artificial intelligence that deals with the analysis, understanding, and generation of natural language. Its functions include sentiment analysis, part-of-speech tagging, named-entity recognition, and speech recognition. Natural Language Processing doesn't have a German Wikipedia article 😉. 👨🎓 "𝐓𝐡𝐞 𝐨𝐭𝐡𝐞𝐫 𝐍𝐋𝐏" 𝐬𝐭𝐚𝐧𝐝𝐬 𝐟𝐨𝐫 𝐍𝐞𝐮𝐫𝐨-𝐋𝐢𝐧𝐠𝐮𝐢𝐬𝐭𝐢𝐜 𝐏𝐫𝐨𝐠𝐫𝐚𝐦𝐦𝐢𝐧𝐠. It describes a collection of methods and communication techniques aimed at influencing psychological processes in humans. This NLP is predominantly practiced outside of academic science and is rejected as unscientific in many circles. Neither I nor Eloquest are familiar with the latter 😉.

𝐓𝐡𝐞 7 𝐃𝐞𝐚𝐝𝐥𝐲 𝐒𝐢𝐧𝐬 𝐨𝐟 𝐌𝐞𝐝𝐢𝐜𝐚𝐥 𝐖𝐫𝐢𝐭𝐢𝐧𝐠 𝐀𝐮𝐭𝐨𝐦𝐚𝐭𝐢𝐨𝐧 👻 Want to speed up your writing process? Good idea! But try to avoid the following common mistakes. ☠ 𝐒𝐭𝐚𝐫𝐭𝐢𝐧𝐠 𝐰𝐢𝐭𝐡 𝐭𝐡𝐞 𝐬𝐨𝐥𝐮𝐭𝐢𝐨𝐧 𝐢𝐧𝐬𝐭𝐞𝐚𝐝 𝐨𝐟 𝐭𝐡𝐞 𝐩𝐫𝐨𝐛𝐥𝐞𝐦. Wondering if there's a more modern way to organize AE tables than plowing through dozens of pages with a highlighter? Randomly throwing a language model at it will just waste your time. Better start with a detailed description of your problem. This will give you a basis for communication with experts. ☠ 𝐑𝐞𝐢𝐧𝐯𝐞𝐧𝐭𝐢𝐧𝐠 𝐭𝐡𝐞 𝐰𝐡𝐞𝐞𝐥. There are a number of companies that have been dealing with topics like the automation of clinical study reports, CSRs, for years: Yseop, Narrativa Generative AI, and of course, we at Eloquest, to name just a few. Get in touch and see what's already available or who could help you with their expertise. ☠ 𝐁𝐥𝐢𝐧𝐝 𝐭𝐫𝐮𝐬𝐭 𝐢𝐧 𝐀𝐈 𝐞𝐱𝐩𝐞𝐫𝐭𝐬 𝐰𝐡𝐨 𝐡𝐚𝐯𝐞 𝐧𝐨 𝐢𝐝𝐞𝐚 𝐚𝐛𝐨𝐮𝐭 𝐲𝐨𝐮𝐫 𝐢𝐧𝐝𝐮𝐬𝐭𝐫𝐲. Unfortunately, medical writing is too complex for that. ☠ 𝐈𝐠𝐧𝐨𝐫𝐢𝐧𝐠 𝐭𝐡𝐞 𝐞𝐱𝐩𝐞𝐫𝐢𝐞𝐧𝐜𝐞 𝐨𝐟 𝐲𝐨𝐮𝐫 𝐞𝐦𝐩𝐥𝐨𝐲𝐞𝐞𝐬. This not only scares your employees but also harms software development. You can't afford developers to develop past the users. ☠ 𝐁𝐥𝐢𝐧𝐝 𝐭𝐫𝐮𝐬𝐭 𝐢𝐧 𝐟𝐫𝐞𝐞𝐥𝐲 𝐚𝐯𝐚𝐢𝐥𝐚𝐛𝐥𝐞 𝐭𝐨𝐨𝐥𝐬. It goes without saying. ☠ 𝐀𝐥𝐰𝐚𝐲𝐬 𝐚𝐬𝐬𝐮𝐦𝐢𝐧𝐠 𝐭𝐡𝐚𝐭 𝐞𝐯𝐞𝐫𝐲𝐭𝐡𝐢𝐧𝐠 𝐲𝐨𝐮 𝐧𝐞𝐞𝐝 𝐢𝐬 𝐢𝐧𝐡𝐞𝐫𝐞𝐧𝐭 𝐢𝐧 𝐲𝐨𝐮𝐫 𝐝𝐚𝐭𝐚. Sorry to break it to you, but the chances are good that your dataset is not comprehensive enough to fully represent the affected document type. Business logic and internal agreements are all too often not documented in the input data. ☠ 𝐒𝐭𝐚𝐫𝐭𝐢𝐧𝐠 𝐚 𝐫𝐞𝐯𝐨𝐥𝐮𝐭𝐢𝐨𝐧. 𝐓𝐡𝐞 𝐝𝐞𝐭𝐚𝐢𝐥𝐬 𝐜𝐚𝐧 𝐰𝐚𝐢𝐭. It may be that your automation project has the potential to revolutionize the medical writing industry in the long run. But to reap the benefits in the near future, don't lose contact to planet earth.

"𝐈 𝐚𝐦 𝐚 𝐦𝐞𝐝𝐢𝐜𝐚𝐥 𝐰𝐫𝐢𝐭𝐞𝐫 𝐚𝐧𝐝 𝐝𝐨𝐧'𝐭 𝐰𝐚𝐧𝐭 𝐭𝐨 𝐦𝐢𝐬𝐬 𝐭𝐡𝐞 𝐀𝐈 𝐭𝐫𝐚𝐢𝐧. 𝐖𝐡𝐚𝐭 𝐝𝐨 𝐈 𝐧𝐞𝐞𝐝 𝐭𝐨 𝐝𝐨?" First of all, it's worth a lot that you've noticed the train. And I can reassure you in advance: It's neither new nor should it cause fear. The development of #GenAI has progressed rapidly over the past two years, but it doesn't replace the job of a medical writer. Why? Because #LLMs predict words, but medical writing is not just about words. Depending on the type of document, it's about what needs to be highlighted. Are they numerical relationships? Does a situation need to be simplified? Does the original text need grammatical changes? Some of these sub-steps have nothing to do with #AI but are simple digitization measures that contribute to the automation of the writing process and thus speed up your work. But back to the original question: What do I need to do? First, ask questions:   🚀 1: Analyze your workflow: Which sub-steps are standardized, which are not? 🚀 2: Business overview: Which sub-steps take up the most time and are therefore the biggest cost drivers? 🚀 3: Which sub-steps are the most nerve-wracking and thus potential employee deterrents? 🚀 4: Demographic change will lead to the retirement of experts in the coming years. Has the knowledge of this age cohort already been passed on to the next generation? For point 2, it's worth seeking expert advice first. Are there ready-made automation tools, and if not, how much would the development of a support system cost at this point? Point 1 will be directly queried by these tools. Points 3 and 4 influence the prioritization of individual automation steps. I hope I could give you a little more confidence on your journey. If you have any questions about specific types of documents or suspect automation potential in your company already: At Eloquest, we also regularly offer workshops on this topic, even within companies.

𝐆𝐞𝐧𝐭𝐥𝐞 𝐑𝐞𝐦𝐢𝐧𝐝𝐞𝐫: 𝐈𝐭’𝐬 𝐀𝐋𝐖𝐀𝐘𝐒 𝐚𝐛𝐨𝐮𝐭 𝐭𝐡𝐞 𝐭𝐞𝐚𝐦! About two years ago, when I was still part of L-One Systems, I gave a talk entitled "𝐒𝐭𝐚𝐫𝐭𝐢𝐧𝐠 𝐚𝐧 𝐍𝐋𝐏 𝐏𝐫𝐨𝐣𝐞𝐜𝐭 - 𝐃𝐨𝐬 𝐚𝐧𝐝 𝐃𝐨𝐧'𝐭𝐬" (link in the comments) in the context of the 𝐀𝐈 𝐀𝐜𝐚𝐝𝐞𝐦𝐲 by hessian.AI. In sum, the #NLP industry has been significantly shaken up over the last two years, which is why I revisited my own talk to see how well it has aged and which take-home messages I would still sign today 😅 . – Fortunately all of them 😉. So much seems not to have changed in the practical use of NLP tools 😉. I would, however, like to explicitly address one of the points: 𝐓𝐡𝐞 𝐭𝐞𝐚𝐦 𝐜𝐨𝐦𝐩𝐨𝐬𝐢𝐭𝐢𝐨𝐧! In other contexts, this is absurdly always clear to us. A world-class goalkeeper will never become a world champion without world-class forwards. A choreography wouldn’t work if the dancers didn’t align. In AI projects, however, it is similar and not all roles must ideally filled with technical expertise. Are the partial tasks staffed with the right people? 👍 𝐁𝐮𝐬𝐢𝐧𝐞𝐬𝐬 𝐈𝐧𝐭𝐞𝐥𝐥𝐢𝐠𝐞𝐧𝐜𝐞: You want to automate a process. Are the experts who currently perform the process manually on board and able to formulate requirements for the software to be developed? 👍 𝐓𝐫𝐚𝐢𝐧𝐢𝐧𝐠 𝐝𝐚𝐭𝐚: Are the data processed? If not, this step still needs to be done. ML engineers are often neither motivated nor trained here. Depending on the project, linguists, lawyers, medical professionals, or graduates of digital humanities programs can help here. 👍 𝐒𝐨𝐟𝐭𝐰𝐚𝐫𝐞 𝐝𝐞𝐯𝐞𝐥𝐨𝐩𝐦𝐞𝐧𝐭: The ML engineer, whom you probably first focused on, does not have to be an expert in usability and GUIs. Are all essential positions filled with the right personae? 👍 𝐂𝐨𝐨𝐫𝐝𝐢𝐧𝐚𝐭𝐢𝐨𝐧: Are the employees responsible for BI, data, and development able to communicate with each other or do they need a mediator (for example, from management)? In order for the show to run, all cogs must be coordinated with each other! 📸 by Eloquest  

𝐈𝐬 𝐭𝐡𝐞𝐫𝐞 𝐜𝐮𝐫𝐫𝐞𝐧𝐭𝐥𝐲 𝐚𝐧𝐲 𝐬𝐨𝐟𝐭𝐰𝐚𝐫𝐞 𝐬𝐞𝐫𝐯𝐢𝐜𝐞 𝐩𝐫𝐨𝐯𝐢𝐝𝐞𝐫 𝐭𝐡𝐚𝐭 𝐝𝐨𝐞𝐬𝐧'𝐭 𝐜𝐥𝐚𝐢𝐦 𝐭𝐨 𝐛𝐞 𝐝𝐨𝐢𝐧𝐠 "𝐬𝐨𝐦𝐞𝐭𝐡𝐢𝐧𝐠 𝐰𝐢𝐭𝐡 #𝐀𝐈"? 𝐀𝐫𝐞 𝐭𝐡𝐞𝐲 𝐦𝐚𝐤𝐢𝐧𝐠 𝐩𝐫𝐨𝐟𝐢𝐭𝐬?   Frankly, the current coverage is getting on my nerves once again. We have digitalization deficits everywhere. Redundant paperwork that could have been automated 15 years ago is still being done manually, but because "software" and "automation" sound too much like "steam engines" and "fax machines", we prefer to discuss the ethical consequences of unborn chicks. Anything else would be too technical (“What a bunch of nerds. I don't understand anyway...”). I would be enormously pleased if, instead of cramming the word "AI" into every IT event title, we would differentiate the topics a little more clearly.   💡 "AI" is not synonymous with "language model." 💡 An assistance system doesn't necessarily have to learn independently to be useful. Not every software has to pretend to have AI in it to have a reason for existence. 💡 We must naturally discuss which regulations we will need in the future. Also, outside of expert circles. However, many discussion rounds currently dealing with "AI" are strictly speaking about digitalization. 💡 Just because natural language (as opposed to programming languages) is involved, it doesn't necessarily mean there's "AI" in it. This point concerns us, Eloquest, as an #NLP company especially. If an assistance system saves several days of manual work because it eliminates tedious copy-paste steps, it is a sensible acquisition from a business perspective, but it does not necessarily mean it's a self-learning system or one that requires an #LLM. What do you think? Can you still manage without AI-washing in your company? Would love to hear your thoughts, Andreas Wieland, Alexander Stumpf, Julia Knorr, and everybody else who feels addressed. For a long time, I thought #ChatGPT could be a driver for societal IT education, but I increasingly perceive it as a blocker. 📸 by TU Darmstadt Lehrstuhl Wirtschaftsinformatik