Sie ist eine der wichtigsten Figuren in der Entwicklung von künstlicher Intelligenz. 10 Zitate von Mira Murati zeigen, wie die Verantwortliche für die aktuell berühmtesten KI-Systeme ChatGPT und DALL-E tickt.
In ihrer Rolle als CTO (Chief Technology Officer) des US-Unternehmens Open AI verantwortet Murati die am schnellsten wachsenden Anwendungen der letzten Jahre. Schon nach zwei Monaten hatte ChatGPT 100 Millionen aktive Nutzer. Instagram dagegen brauchte dafür zweieinhalb Jahre.
Die Faszination für die Systeme, die Bilder und Texte (Storys!) selbsttätig auf Basis einfacher Vorgaben oder Fragen entwickeln, ist riesig. Natürlich gibt es auch Unsicherheit. Mira Murati betont, dass das Feedback menschlicher Nutzer zentral sei, um die künstliche Intelligenz einzunorden und vor Missbrauch zu schützen. Wir alle sollten ihrer Meinung nach diesen technologischen Wandel mitgestalten.
Die Zitate von Mira Murati habe ich aus Interviews und Diskussionen gefiltert. Für mich zeigen sie die Klugheit und die Umsicht, mit der diese radikal neue Technologie entwickelt wird. Das Spannende an Murati ist für mich, dass sie eine Macherin ist, die sich äußert. Sie weiß exakt, wovon sie spricht. Ihr Storytelling ist auf hohem Niveau. Es ist sachlich-beschreibend, klar, schnörkellos und trotz ihrer hohen fachlichen Expertise sehr gut für Laien verständlich.
Zitate von Mira Murati:
Vision? „We are trying to teach machines to see the world in a similar way that humans do.“
Lernen mit KI? „It has the potential to really revolutionize the way we learn. People are in classrooms of, say, 30 people. Everyone has different backgrounds, ways of learning, and everyone is getting basically the same curriculum. With tools like ChatGPT, you can endlessly converse with a model to understand a concept in a way that is catered to your level of understanding. It has immense potential to help us with personalized education.“
Jobkiller KI? „I do think as with other revolutions we’ve gone through there will be new jobs and some jobs will be lost and there will be some retraining required as well. But I’m optimistic.“
Hat KI eine Persönlichkeit? „These Models don’t have a specific personality or opinion, they are just based on data we trained them on.“
Wie kommen wir zu einer KI-Ethik? “This is a unique moment in time where we do have agency in how AI shapes society. And it goes both ways: The technology shapes us and we shape it. There are also a ton of questions around societal impact. And there are a lot of ethical and philosophical questions that we need to consider. And it’s important that we bring in different voices, like philosophers, social scientists, artists, and people from the humanities.”
KI demokratisiert die Kreativität
Fortschritt? “You could make technological progress in a vacuum without real-world contacts. But then the question is, are you actually moving in the right direction?”
Positiver gesellschaftlicher Impact? „An AGI system (künstliche allgemeine Intelligenz, vergleichbar mit der menschlichen Intelligenz) is certainly bound to have a big impact in our lives, in the world. And I think it can have a very positive impact because there are few challenges today in the world that we cannot solve on our own. I don’t see how we can solve them without technology, such as climate change and disease. And I think it can help us if we apply care and foresight in the development and the deployment of the systems.“
Hat KI ein Bewusstsein? „It is hard to determine whether a non-biological system has the emotional or spiritual experience that we have, even if the system can claim that. There is not scientific way, at least so far, to state objectively that the system is conscious or not.“
ChatGPT und DALL-E sind Tools
Warum lernt KI Sprache? „With language in particular we know that there is a lot of data in natural language available. Some consider understanding language as the pinnacle of intelligence. Having a shared understanding of context and being able to reason. Some people believe that if you can really master language then you have an intelligent system. Truly our goal is to build this general intelligence system. And our hypothesis is that if we can get them to see and understand the world in a way that is similar to humans, perhaps we can get there.“
Storytelling? „There isn’t anything particularly new about having a human helper. These concept of extending the human abilities and also being aware of the vulnerabilities are timeless. And in a sense we are continuing this conversation by building AI technologies today. We see GPT and DALL-E as tools, extensions of our creativity or writing abilities.“
Ist menschliche Kreativität nichts mehr wert? „Technologies like DALL-E can democratize high-quality creation of images and ideas and it can push them so much further. We are often asked: Does that dilute the human creation? I think, maybe there is going to be a more nuanced appreciation for this co-creation and also for the original human creation. For example, today you can have a skilled artist that went to art school that can try to create a replica of the Nightwatch (Rembrandt). It might look great but we will still appreciate the original painting differently.