SXSW Revelations of Generative AI
With the European AI week almost ending, we have to talk about what happened at SXSW. Every year, tens of thousands of media enthusiasts travel to Austin, Texas, to see and participate in all the activities that are happening in these 10 days. This year, Future Media Hubs was present together with many of its partners and the talk of the day, every day, was on AI. We’ve gathered the most important insights to bring them to you.
The Internet is Dead
According to many, Amy Webb was the highlight of the festival. On the morning of March 11th, people were standing in line at 8 AM to get a peak into this featured session. And their enthusiasm did not disappear during her keynote. Amy Webb is a futurist, founder and CEO of the Future Today Institute and professor at NYU Stern School of Business. During her keynote and in the 2023 Emerging Tech Trend Report, Webb tries to tackle 35 trends and 300 slides in 60 minutes. She touches upon many different sectors, but in this article, we’ll focus on Artificial Intelligence.
Amy Webb starts the AI trend with a bold statement: “The internet as we know it is over.” In a 2017 Google Report, the concept of Transformers (neural networks that can learn context in sequential data) was introduced. These underpin large language models (LLMs) who can recognize, summarise, predict and generate text and other types of content. While Transformers work like our brain, LLMs work like us, humans, which, according to Amy Webb, leads to a pessimistic view where everything becomes information and nothing will be found. Next up is criticism about ChatGPT and the dangerous security and transparency issues and suspicions that surround it. She gives the example of OpenAI, who in 2020 declared themselves that GPT-2 was “too dangerous to publish”. Next to this, the computing power needed to develop new AI models is in the hands of the clouds of AWS, Google or Microsoft.
She then goes on and launches the great AISMOSIS: “the moment when AI systems interact with all of the data through any source or situation. And once that transition happens, we’re going to be living in a different world”. In the future, everything in our lives, all systems and organisations, will be increasingly interconnected with AI. According to Amy Webb, this is the next step in evolution, but we need to be aware that there is a high probability that it will affect our lives and happiness in a negative way. For example, when she asked Generative AI for an image of her conference at SXSW, a white male vision was generated: A white man talking about AI to an audience of other white men. Ian Beacraft, CEO of Signal And Cipher, adds that it is not only about creating, but also about building a relationship with AI. In his vision, people will not even be able to tell the difference between human and AI interactions. He then shows a short video of actor Keanu Reeves explaining the plot of the Matrix movie to a 13 year old girl. She simply tells him that it doesn’t even matter what is true or not. Sounds like ethical problems are just waiting around the corner?
OpenAI
It seems like Amy Webb is not the biggest fan of Generative AI. Greg Brockman, Co-founder of OpenAI and one of the biggest tech-stars at SXSW, begs to differ in his featured session on ‘ChatGPT, DALL-E and the Impact of Generative AI’. At the start of the session we take a step back and look at the launch of GPT3 (or 3.5 as he likes to call it) in November 2022. Brockman states their biggest takeaway is that OpenAI informs people about what is possible. He then goes on and counters Amy Webb’s doubts about the different ChatGPT risks by explaining they’ve elaborately tested and anticipated for most of the negative and disruptive use cases before launching.
With an anecdote about an illness his wife had, Brockman explains that the intention of ChatGPT in medicine is to give you an idea of what type of illness you could have, without setting doctors on the side completely. OpenAI should suggest things, open up your mind and be an assistant that is willing to do whatever you want, 24/7. For example, Generative AI can function as a helpful brainstorm partner to accelerate the generation of new ideas. Both Amy Webb and Greg Brockman agree that AI will not take over our jobs, but just make them more efficient. AI is an amplification of what humans can do. It will be the assistant to us being the director and in this point of view, every aspect of our life will be amplified. In terms of media, Generative AI can help create more personalised experiences such as alternative endings of TV shows or writing yourself into a storyline. We’ll leave it up to you, our dearest reader, to decide if you want to rewrite the ending of Game of Thrones. But for content moderators, the future does not look as bright as pictured. Greg Brockman states that they are most at risk since you don’t want a ‘human influence’ in content moderation anyways. We’ve followed a session on this topic called ‘Evolving Content Moderation: The Next 10 Years’ and even though it was really interesting, we were left behind with the question “Well, what is actually going to happen?” We hope to find out what their idea is and why they did not discuss AI in one of the upcoming articles.
Painting the Future
According to the co-founder of OpenAI, ChatGPT will be the new frontier for most areas. The barriers to enter will decrease as we’ve seen with DALL-E. But what about ethics? For Greg Brockman, we should take it to the public, implicating that he doesn’t have the answers. He broadly poses a question about how to make sure everybody rightly benefits from the use of AI, but remains unclear on how to actually do it. In another session ‘Deepfakes to DALL-E: Real Rules for Fake Media’, Andy Parsons, Senior Director of Adobe’s Content Authenticity Initiative, also remarks “We talk a lot about responsible AI, but we never actually say who is responsible.” Remember Amy Webb’s questions about the different models AI is based on and the cloud computing power necessary for AI? Vague non-answers were given here as well by Brockman. Webb’s sceptical vision also reminds us that we have to be aware of a possible new and even more dangerous digital divide in society, now about having access to AI.
“People who are not empowered to use new AI tools may soon be completely left behind.”
Amy Webb, founder and CEO of the Future Today Institute
When going on to transparency in AI and really knowing what is real and what not, Brockman remained vague about a solution. He just mentioned that we as humans have always been good at writing fake texts, but the barrieres to do so right now are much lower than in the past. One of our partners followed the session ‘AI in Newsrooms: What’s the Impact on Journalism’ and we would like to give you the two most important takeaways:
- The importance of integrating the right rules and standards into AI models;
- The importance of transparency around the use of AI in newsrooms.
“Transparency is the new objective for journalists.”
Aimee Rinehart, Program Manager for the Associated Press Local News AI initiative
David Smydra, former Head of HITL Curation at Twitter, tells us about the dedicated standards team at Twitter that is responsible for developing the guidelines and standards for responsible curation. He believes this is a critical element for newsrooms as well. Laura Ellis, Head of Technology Forecasting at BBC, explained a framework for synthetic media where certain ‘business rules’ can definitely help the process. For example, machines don’t understand that topics such as suicide are to be handled with care and delicacy, either for legal or harmful reasons. Having these rules when it comes to recommendations can beneficially help to tailor content for specific users.
At Associated Press, they try to be as transparent as possible when an article or report is created by LLMs. But to be transparent, journalists will need the full technical expertise to understand and explain how these technologies are being used. With the upcoming elections, this is an even more prominent debate because how can we make sure that what we hear is actually true and not a deepfake generated to get more votes?
A lot of open questions remain and the biggest one is why Greg Brockman did not announce GPT4 at SXSW? When asked about the new technology they were building, he did not disclose much information. Was it a missed opportunity or a well placed move to avoid critique since it was world news only four days later?
For many, however, one thing is clear about AI: it is here to stay and will play a major role in the future. It will be “a tool, just like the cell phone in your pocket” as Greg Brockman likes to call it. AI will disrupt a lot of industries and will be a big thing in the metaverse, but that is a conversation that we’ll have another time. Stay tuned for next week, where we’ll discuss all things metaverse at SXSW.
