.webp)
Responsible artificial intelligence in relation to generative AI
Advances in the field of artificial intelligence (AI), particularly in the context of generative AI, have opened up completely new fields of action in recent years that were previously unrealizable. The ability to have natural language interactions with AI models has significantly improved the accessibility of AI for society. Even people with no previous experience of AI can now use these models and encounter new application scenarios.
In practice, numerous solutions have already been developed using generative AI models. One example of this is the pharmaceutical company Boehringer Ingelheim, which uses a specially developed knowledge management solution to support its scientists in researching new drugs. The researchers use the Azure OpenAI service to search and compare scientific documents in their database. This enabled them to save an impressive 150,000 working hours within 70 days, which in turn could be invested in drug research.
Guide for AI
Es ist erfreulich zu sehen, dass die breite Öffentlichkeit zunehmend die <a href="https://www.taod.de/ai-potenzialanalyse-mit-data-thinking-workshop" data-webtrackingID="blog_content_link" > Potenziale erkennt, die KI </a> für unsere Gesellschaft bereithält. Allerdings dürfen wir dabei nicht die potenziellen Risiken und Gefahren aus den Augen verlieren, die mit dem Einsatz von KI einhergehen. Im Jahr 2019 hat Microsoft sechs grundlegende Prinzipien zur verantwortungsvollen Entwicklung von KI-Systemen formuliert und einen Leitfaden veröffentlicht, der als Orientierung für Entwickler und Organisationen dient.
These principles include:
- Fairness: AI systems should be fair and non-discriminatory. It is important to ensure that AI applications do not reinforce existing biases or create new ones.
- Reliability and security: The system should function reliably in various application scenarios, including those for which it was not originally intended.
- Data protection and security: The protection of personal data is paramount. AI systems must be developed securely and in compliance with data protection regulations. Data must not be leaked or disclosed to the outside world.
- Inclusion: The system should include people with different abilities. To achieve this, minorities should be involved in the planning, testing and development of AI systems.
- Transparency: Those who develop AI systems should be open about how and why they use AI and communicate the limits of the system. In addition, the functioning of AI models should be transparent. Users should be able to understand how decisions were made.
- Responsibility / Accountability: Developers of AI solutions must be aware of their responsibility and ensure that AI systems are used ethically and responsibly.
Numerous tools and processes have already been developed in the past to support the implementation of these principles. Generative AI is a particularly fascinating technology, as it is able to generate completely new content and thus become creatively active. However, this new field of activity also brings with it a number of challenges, particularly with regard to the implementation of the aforementioned principles.
Generative AI is a particularly fascinating technology.
RAG against hallucinations
One of these concerns the non-determinism property of this technology. Generative AI models provide information without confidence scores, a traditional method for evaluating the results of machine learning models. This changes the steps to ensure the principles of transparency and fairness.
When a large language model makes incorrect statements, we speak of "hallucinations". The model reproduces information that sounds plausible and comprehensible at first glance. The concept of "Retrieval Augmented Generation" (RAG) can be used to counteract this problem. With RAG, the existing knowledge of the model is not used. Instead, only the model's ability to understand and generate natural language is used.
The approach combines the best of two worlds: Information search and creative text generation. The process begins with the search for information. This involves searching for existing information that matches the specific query. This "raw material" can come from a database, a text corpus or even the Internet. It is then generated. Once relevant information has been found, a creative approach is used to generate a coherent answer. This involves adding your own words, structuring the text and adapting it to the context.
Active process
In der Tat sind mit der Entwicklung generativer KI neue Herausforderungen im Bereich der verantwortungsvollen KI entstanden. Gleichzeitig eröffnen sich jedoch auch vielfältige Chancen. Die Möglichkeit, in natürlicher Sprache mit generativen KI-Modellen zu interagieren, macht KI für alle zugänglicher, die zuvor keine direkten Berührungspunkte mit dieser Technologie hatten. Immer mehr Menschen können nun KI nutzen und Teil von Entwicklungsteams für <a href="https://www.taod.de/services/artificial-intelligence" data-webtrackingID="blog_content_link" > KI-Lösungen </a>werden, die auf fundiertem Anwendungswissen basieren, weniger auf rein technologischem Know-how.
Another advantage is that AI is also accessible to people without many years of programming experience. By generating code using GitHub Copilot, they can create program code or better understand existing code. AI is no longer a niche topic for technology experts, but a relevant and tangible field for everyone.
Responsible use of AI is an ongoing process.
Abschließend möchte ich betonen, dass der verantwortungsvolle Umgang mit KI ein kontinuierlicher Prozess ist, an dem Unternehmen aktiv mitwirken sollten. Da sich <a href="https://www.taod.de/company/partner-technologies" data-webtrackingID="blog_content_link" > Technologien </a> ständig weiterentwickeln, müssen auch unsere Kontrollmechanismen entsprechend erweitert und angepasst werden. Wir sollten neue Technologien mit offenen Armen begrüßen und gleichzeitig sicherstellen, dass sie verantwortungsbewusst eingesetzt werden. Nur so können wir das volle Potenzial der KI nutzen und gleichzeitig ethische Standards wahren.
Dieser Artikel erschien in ähnlicher Form erstmalig in <a href="https://www.taod.de/data-magazin" data-webtrackingID="blog_content_link" > Ausgabe 01/24 </a> von data! Alle Ausgaben und Artikel unseres halbjährlich erscheinenden Magazins findest du hier:
<a href="https://www.taod.de/data-magazin" data-webtrackingID="blog_content_link" > data! Magazin: Cloud Services, Data Analytics & AI | taod </a>
Would you like to implement AI in your company?
About Sophia Cullen
Sophia Cullen is a Solution Architect for Artificial Intelligence at Microsoft. She works closely with Microsoft partners to implement AI projects for joint customers. With extensive experience in this area, she has successfully managed numerous projects. Her primary goal is to encourage customers to explore new fields of action and identify innovative solutions. As a member of the Responsible AI V team at Microsoft, she helps to raise awareness of ethically responsible AI within the organization.
%20(1).webp)
