The Future of Fakery: Deepfakes, Generative AI & The Fight for Authenticity

Talk
This event has now finished.
  • Date and time: Friday 17 November 2023, 6pm to 7.30pm
  • Location: In-person and online
    K/133, King's Manor, Exhibition Square (Map)
  • Audience: Open to staff, students, the public
  • Admission: Free admission, booking required

Event details

In this presentation, Bo Bergstedt will explore the expanding realm of Generative AI, focusing on its applications in image, video and audio generation. He will delve into the state of deepfakes in 2023, technical advancements and illustrating their implications. The talk will address the challenges posed by generative AI for the media and news industry, possible consequences and ask you to think about what might be next.

View the recording of this event on our YouTube channel

Q&A transcript

Q: Which sources/people on twitter or linked in should one follow

A: I am so not a social media person, so I don’t follow a lot. I rely on subreddits on Reddit that usually brings the most relevant stuff to the frontpage. On Reddit there are subreddits dedicated to OpenAI, ChatGPT, Dall-E, Midjourney, Stable Diffusion, Machine Learning etc. On LinkedIn there are many people sharing the same stories and insights on AI every day. David Caswell you had as guest is good at having a focus on News – and another is Ezra Eeman who creates a popular newsletter called Wayfinder every week revolving around News and AI. On Twitter/X I guess following the CTOs, CEOs and Head of Science from the different tech companies would make sense. Its however very often either superficial or extremely “technical”.

Q: Can you recommend any training courses? Events for people who want to start using AI?

A: Google, Microsoft and Amazon are all providing free courses to understand Generative AI. They vary in how technical they are and of course they have an interest in telling their version with focus on using their tools/cloud – but still a great source of information and learning But there are tons of great courses and content for learning available from universities, organizations and companies. Coursera has gathered most of it here – and on the left side you can choose your level (beginner, advanced etc.) and any special interests. 

Q: Suggestions on which tools to maybe get started with?

  • Text/Language: ChatGPT (paid version is a lot better than the free tier), Google Bard and Microsoft/Bing CoPilot (find your own preference – but ChatGPT is ahead currently).
  • Images: DallE3 (either through Microsoft Image Creator or ChatGPT), Midjourney or Stable Diffusion if you want to dive in at the deep end.
  • Video: RunwayML, Synthesia or PikaLabs
  • Voices: ElevanLabs (voice-cloning)

 Q: Tips around AI and HE and again where to get started to stay on top of it?

  • Same as first question - Reddit 

About the speaker

Bo Bergstedt

With a career dedicated to product development in the media industry, Bo Bergstedt has transitioned into a role as a Technology Strategist. In this capacity, he is tasked with following technological trends and their potential impact on danish public broadcaster TV 2, identifying both opportunities and challenges that may arise.

In recent years, this means closely monitoring developments in blockchain, gamification, streaming and the Metaverse. However, primary focus has centred on Generative AI, a field Bo ventured into in 2018 with the creation of deepfakes and synthetic content. Since then, generative AI has been utilised in dedicated projects to create both videos, images and voices at TV 2.

Partners

University of York - Barclays Eagle Labs

Contact

Nadine Waehning