Last week, Jakub Tomczak shared his insights with us during the AI Meet-up Summer Edition.
He introduced us the key components of generative modeling and presented success stories and developments of generative AI. And while it has shown us that almost everything is possible, it is also raising challenges. Resident Romain Huet (ReSleeve) shares his takeaways from the presentation with us:
Trustworthiness: As we embrace tools like ChatGPT, ensuring their trustworthiness becomes paramount. While AI advancements bring unprecedented possibilities, they also raise the specter of fake news and deep fakes in voice, video, and images, which can have far-reaching consequences.
More impressions on our Flickr page.
Sustainability: In the face of the pressing climate crisis, we must carefully consider the environmental impact of generative AI models. For instance, we should explore lean language models, instead of large language models, with fewer parameters and adopting more efficient ways of training.
Education: Since its release, ChatGPT has experienced widespread adoption. Its role in education, however, remains a topic of contention, as it could be a threat to the reputations of knowledge institutions. Universities should look into how to redesign their courses and the way of teaching, so that people actually should use ChatGPT, along with other means.
The challenges we face also give rise to numerous research questions. In navigating the opportunities and challenges of generative AI, we must discuss regulations. Not only to ensure trustworthiness but also to protect artists’ rights. Together, we can harness the potential of generative AI for positive impact.
A big thanks to Jakub and Romain for sharing their insights!
Would you like to be part of our next event?