Putting The Fizz into Transformative Futures

As part of the Sustainability Accelerator's ongoing research into The Rift, we are working with several groups to understand what kinds of practices can most effectively help us to navigate emerging conditions of climatic and societal disruption.

As a practice, futures seeks to understand the forces that shape the future, so that we create better choices in the present. And as we navigate the first years of our new reality–our overlapping crises of climate, biodiversity, economics, energy and geopolitics– we've been exploring what role innovative approaches to futures might have to play.

This has led us to think about how AI can be used well, to augment human abilities and spark the making of new meaning. One outcome is The Fizz, an AI-assisted futures tool, which we have developed in collaboration with eminent futurist Andrew Curry and strategist David Bent of Atelier of What's Next.

In this piece, Andrew outlines the underlying approach to thinking about the future that is embodied in The Fizz–and how it helps us to understand and respond to our current moment.

___

Ideas about the future aren’t neutral. They make constant claims on our time and our attention, often without our realising it. These ideas are designed to influence the future that we get. So it does matter who has the right to engage in conversations about the future, who gets this opportunity, and the approach used to generate those futures.

The quality and breadth of our futures conversations now shapes how the world will become. This is not a new idea. 60 years ago the Austrian activist Robert Jungk wrote, with Johann Galtung, “The future belongs to all of us, not only to small oligarchic groups and interests”. Good futures work should help us to build more resilient societies with the capacity to flourish even in the face of adversity.

But, as Sarah Houle puts it, “We face a significant problem: we’re lacking new futures to pin our hopes on.” Instead, our futures discourse is dominated by the dreams of billionaires, who seem to have lifted them wholesale from the comic books of their childhood: going to Mars, conscious machines, endless lifespans. Or, in the evocative title of a recent book, ‘More everything forever’.

Doing any kind of futures work is difficult. At one level it is an innate human capacity—our brains, according to some recent research, are engaged in mental time travel almost all of the time. As Dean Buonomano explains in Your Brain is a Time Machine, we access memories, “using these memories to conjure future scenarios, understanding the difference between the past and the future, and the ability to judge whether the simulated future is desirable or not.” This is, however, hard cognitive work. Yet building cognition, both individual and shared, is essential to any meaningful futures process: as futurists like to say, “the process is the product.” New futures emerge through this social construction.

And into this mix come Large Language Models (LLMs) and Artificial Intelligence (AI), whether we like it or not. Rather than complaining about this, we need to find ways for LLMs and AI to help make futures thinking more effective, rather than replicating – or growing – poor practice.

Much of mainstream futures practice emerged during the 1970s and 1980s, as a way for large institutions to maintain themselves in the face of a succession of external shocks. This is seen in its methods. Scanning – the process of looking for early signs of change to understand their potential future impacts – is dominated by external, structural, large-scale and easy-to-spot “drivers of change”, which tend, if used uncritically, to reinforce mainstream perspectives.

Scenarios methods–which build a set of multiple possible futures–and, in particular, the 2x2 “intuitive logics” framework, both reinforce this and can erase conflicts of values and ethics from the futures discussion. This is one reason why Richard Slaughter described such approaches as “flatland”. The “decision focus” at the start of the 2x2 process precludes such large questions. If some might escape, the notion of “plausible futures” is often a code for “not in my worldview”. As Shirin Elahi writes, institutions avoid discomfort by erring “on the side of familiarity and uncertainty denial.”

But even when all of these issues are taken into account, there are still multiple barriers to good futures work, especially if you care about who gets to be involved in our futures conversations. Horizon scanning is cumbersome and requires some expertise, and therefore requires resources. It can also be hard for people to know what to do with futures material when they see it.

On top of this, people can be weighed down by the present; they can have a failure of their futures imagination. All of this can get in the way of creating a shared process that is designed to build shared meaning in the service of ‘better futures’.

By ‘better’, we mean futures practices that are open to more and different voices, and which liberate the imagination while respecting the real constraints faced by humans. Better futures methods should also be more effective at creating constituencies for change, through dialogue and social practice. Put simply, new ideas about the future that have the potential and the energy to create change emerge only through interaction. People using ‘better’ futures adjust their assumptions and norms; they make different decisions.

But there is also scope to liberate futures from the technical and analytical space it often inhabits, to engage more with culture, and to accelerate the pace of futures learning. This is why we set out to build the Fizz – a tool to disrupt traditional futures practices. Given the scale of the challenges that humankind currently faces, we need to broaden our discourse about the future and to create a bigger range of ideas about what could be possible.

In developing The Fizz, this broad set of conclusions evolved over time, through an extended process of meetings and workshops. But in effect, these conversations created a set of problem statements that we set out to solve:

• How do we create richer futures conversations to stimulate better decision-making?
• How can we get more quickly to effective futures conversations?
• How do we extend people’s futures imaginations?
• In doing these things, what can LLMs and AIs do that is different from what people do?

Some of the answers to these questions already lie in the futures literature. For example, The Fizz (influenced by the Seeds of a Good Anthropocene project) focuses on weak signals rather than the more familiar and more mainstream ‘drivers of change’ that tend to underpin scenarios work. Weak signals typically describe edge behaviour by people who have a dissonant view of society.

At this stage in its development, we have been using the LLMs in two ways. First, a scanning engine that picks out ‘seeds’, those weak signals that are relevant to the Rift research programme. Second, a generator engine that creates “futures packs”– narrative elements of a future in which input seeds have become dominant, era-shaping phenomena. But the packs are only prompts for conversation, and these are deliberately incomplete. Rather than being polished scenarios, they are overlapping perspectives on a possible future world. To be interactive, it also needs to be unfinished, as Brian Eno has written – inviting human users and interpreters to work with the gaps.

These fractions of the future are also designed to stimulate the imagination. They can include radio bulletins or stories in different styles, to help people gain a foothold in the future. Building on the work of Joseph Rowntree Foundation, our collaborator David Bent has described The Fizz as “an imagination infrastructure”, in embracing an approach that is critical of existing social relations, creative in exploring new possibilities, and empathetic towards multiple perspectives.

But that’s enough about humans. What we’ve discovered as we have worked with The Fizz, which is still in its early stages of development, is that because the AI is continuously scanning, it brings into its descriptions of its future fragments a wider range of ideas and concepts than a group of humans would identify.

For example, in a recent test run on the future of digital technologies, it described a future in which monthly “Brownout Drills”, designed to ensure that backup systems, which used a combination of paper chits and legacy mainframe software, are signed off with a tamper-proof graphene ink. In conventional futures approaches it can take people weeks of immersion and multiple workshops to get to this kind of ‘what if’ starting point.

___

There is no shortage of ideas about the future out there, but a lot of them are “used futures” that have little to say about the actually existing world we find ourselves living in at the moment. But there is a shortage of imaginative futures that help us to stretch our sense of what might be possible. The Fizz is a tool to help stretch the futures imagination. It is designed to help us to create new and different routes through the crisis, in the hope that more visionary choices might lead us to transformational futures rather than to collapse. We’ll see you there.