Generative AI: the science experiment of our times? 

It can be hard to separate hype from reality in all the headline generative AI systems, such as ChatGPT, have been making of late. Inside ICE asked Dr Maya Indira Ganesh and Dr Jonnie Penn, Co-Directors of ICE’s award-winning, part-time Master’s in AI Ethics and Society, to give us their takes on where the tech’s at right now. 

There are so many stories swirling around generative AI, including the AI-driven chatbot, ChatGPT. What is it? 

Jonnie: Generative AI uses statistics to replicate images, text, audio, and so on based on training data. The degree to which this output is creative, or derivative is strongly contested. 

Maya: It’s statistics on steroids. The word ‘generative’ does not specify much beyond acknowledging that it’s generating things. It’s up to the user to do things with it. 

Jonnie: It helps to consider generative AI’s output, which is synthetic media. For example, seeing an actor’s ‘face’ and hearing their ‘voice’ without the actor’s involvement. These tools are becoming affordable and available, which presents exciting possibilities for communities like the tech industry, but also challenging consequences for others, like artists.  

What’s exciting about the stage we’re at right now? 

Maya: Somebody described it to me as like a scientist’s lab, where you’re throwing things together to experiment with. Artists, for example, are already playing with this new technology, even with concerns about what it means for their own professions. 

Jonnie: What excites me about generative AI right now is the possibility of reaching a broader consensus about the strengths and limits of digital technologies. The press puts ‘AI’ forward as something that can resolve this or that, but those who know these tools best are, often, ill-equipped to lead an inclusive conversation about whether or not they’re in over their heads.  

Recently, ChatGPT was banned in Italy over privacy concerns and AI experts sent an open letter asking governments to pause giant AI experiments until “we are confident that their effects will be positive and risks manageable.” 

Jonnie: I take that letter to be a cry for help among developers, asking for other experts and civic leaders to take the reins. You don’t need to know how an engine works to regulate vehicles, and I think the same is true here, although wealthy companies will demand carte blanche to do what they want. 

Maya: I see a couple of other aspects, too. The biggest signatories are all insiders and elites, so you could view this as a move to encourage self-regulation, rather than let governments take control. As in, “We know this might become problematic, but we can regulate it because we’ve seen the dangers.”  

Another perspective is put forward by Lee Vinsel, a US science and technology scholar, who coined the term ‘criti-hype’: hype-driven faux concern that deliberately pushes technology in a supposedly unwanted direction. Vinsel advocates not paying attention to this hype at all. 

How does the AI Ethics and Society Master’s at ICE navigate these issues? 

Maya: For Jonnie and me as Course Directors, the open letter is a flashpoint that allows us to examine the controversies unfolding right now about the shaping of AI and the powerful social actors leading it. It’s important for us to bring all these perspectives together and show that all of them are real to some extent and must be part of the discussion. 

Jonnie: The course allows people to learn the vocabulary and frameworks about how we integrate something like generative AI without toppling the ship. We say to our students that this is a laboratory, and you’re with us as peers thinking through how to do this work. It’s a collaboration around what the limits might look like for governments, law firms, companies, technologists and so on. 

To find out more about ICE’s MSt in AI Ethics and Society, winner of the CogX Award for Best Course in AI, visit: https://www.ice.cam.ac.uk/course/mst-ai-ethics-and-society