It’s no secret that in recent years, SXSW has become more about what cool tech can do than why it should do it. That said, the absence of a substantive and actionable discussion around digital ethics was more greatly felt this year than years past because of how advanced and immersive the technology has become. By the time the first talk was done, it was clear that this year’s speakers no longer pondered the probability of human-machine symbiosis, but rather, when and how so.
AI, machine learning, and enabling systematic collective thinking in humans
One of the biggest arguments I stumbled across this year during SXSW was around differing models of the man-machine intelligence relationship. Traditional models tend to position humans and machines running in distinctly separate tracks that intersect at key points in time. AI informing human decisions through access to broader, more accurate data sets. Or, humans guiding AI analysis with insights and questions that align to broader human goals. For many, this model simply “feels” safe – human contribution in the future state is assured, as is our authenticity and autonomy.
Newer models (such as those encountered this year), however, present humans and AI fused together in a systematic idea of collective knowledge sharing. Often referred to as “swarm intelligence,” these models blur the line between human and AI contributions by having both humans and AI exist in a single, multidirectional and exponentially expanding grid. Images of the “Borg” from Star Trek come to mind for many; and while human participation is confirmed, human authenticity and autonomy are not.
The debate between these two models is less about their validity and likelihood to exist and more about their inherent ethics. The first model struggles with ensuring that bias is kept out of the algorithms and datasets. The newer models which focus on collective knowledge systems could solve this and create greater empathy because it is inclusive by design, however, these models can seem more daunting from a participation perspective. The burden to ensure that humans are ready to contribute to these new systems will ultimately fall on designers to solve. Unfortunately, this is where the conversation typically stopped.
As exciting as it is to imagine the future that might be, that is exactly how daunting the reality of achieving that future will be. There is an entire mindset change that will need to happen for people to be able to truly realize the value of a new system that values transparency and access for all over competition and hierarchical rewards. While only a few of the speakers took these challenges head on, all were confident that empathy would sit at the heart of not only the question, but also the answer. Coming out of SXSW, the key take-away for this future vision was that there is real work to be done – and it will require aggressive collaboration among us all to move things forward.
Human digital immersion in humane contexts
The other big topic at SXSW this year was about a class of new immersive experiences powered by AI that had the ability to change people’s perceptions of themselves and the world around them. For participants, these experiences feel so real that the line between where the immersion ends and reality “picks back up” can become unclear. They don’t carry with them a simulated experience when they exit immersion; they carry with them a very real and tangible experience that is theirs and will always be theirs. While this degree of immersion can create incredible opportunities for helping people to understand scenarios they otherwise would have no actual exposure to, there are very real concerns that go along with exciting new capability.
Similar to the discussion on collective intelligence, if we are not careful how we create these experiences – we risk losing our sense of self in the mix. In this scenario, however, the man-machine relationship has the ability to alter the perception of what is real, what is not real, and whether it even matters. These experiences also leverage data to make them feel authentic – data that can carry its own biases of the past it is based on. The bias becomes part of the “new” reality, informing the participant’s new understanding of that reality and becoming a part of his or her personal truth.
Absent from the discussion almost entirely was the need for AR/VR experiences to have a clearly articulated intent and desired outcome. If participants are unclear as to whether or not they have left the experience, how are they to know what the purpose was in their having experienced it? The value of the experience for participants is its authenticity – “tricking” the mind into believing its authenticity is what changes them so profoundly. However, this opens the possibilities of manipulation – either real or perceived – that can lead to a lack of trust in what was experienced and a loss of clarity about next steps. Instead of positive actions and behaviors resulting from the experience, emotional trauma can emerge.
While the idea of experiential transparency was sometimes hinted at, it wasn’t often clear if anyone had even a hypothesis about what that might look like and how it would be applied without eroding the impact and value of the experience. In addition to onramps, the experience needs to have clearly marked and smoothly designed off-ramps that transition us back to reality. Off-ramps that ensure we know where we are presently, where we have been (virtually as well as in reality), and where we should go next. One thing that was very clear leaving these sessions: this new crop of digital experiences need to be designed with limits that prevent us from losing our sense of self and experiential history.
In the end, perhaps the most startling thing to say about the AI conversation at SXSW this year, overall, is that there really was not anything new to say. Whether this was for the benefit of the audience – because speakers felt that there was still not enough foundational understanding to move the conversation forward – or, because there was not enough progress made overall, was unclear. But the omission of actionable discussion was felt.
Sessions and resources