The last decade-plus encompasses a rapid and global acceleration within the field of Artificial Intelligence (AI). More specifically, the exponential rise of computational power combined with the wave of big data has yielded some of the most advanced deep learning models to date. What does this mean for our future as a society? It means we are on the path to Artificial General Intelligence (AGI), or the level of AI that is equal to that of human beings (1).
The exact timing of when we will reach AGI is up for debate among experts. Some believe we will reach AGI in as little as a decade, while others think it might not happen until 2100 (2). Regardless of the exact timing - it is clear that through the complex, multi-layer neural networks employed by Deep Learning methods and the growth of data production (in fact - 90% of the world’s data was generated just in the last two-plus years, (3) the path to AGI and possibly artificial superintelligence (ASI) is being laid out right in front of us. In turn, this proliferation of AI technologies along the path to AGI has also resulted in a pseudo arms-race, not only among large tech companies invested and acquiring AI startups but also on a geopolitical level (see President Trump’s “American AI Initiative” or China’s “New Generation Artificial Intelligence Development Plan”).
Why is this so important? It means we - as a global society - need to move from theorizing to actions when it comes to shaping the future of AI for humans. The dialogue around how to use AI for good is already happening. Simply set up a search for either ethical AI, or responsible AI, or transparent AI, and you will see the work that is being done by foundations, NGOs, and even large tech organizations. Since we are on the path to AGI, then these conversations centered around a better AI future need to happen now and be backed by policy and execution. The actions listed below suggest how we are beginning to shape a better AI future and how we should continue to do so. Each item is unique, but not necessarily mutually exclusive of the others.
1. Establish an AI literacy that encourages diversity in thought, dialogue, strategy, and goals.
As mentioned by Jim Stolze at the AI for Good Global Summit in Geneva (hosted by ITU in partnership with the XPRIZE Foundation and ACM) you can’t have a robust conversation about how to shape a better future for AI without AI literacy. In short, if everyone becomes well-versed and educated around the fields of AI and data, then this will lead to a more diverse and inclusive dialogue. Diverse opinions that come from a deeper pool of citizens well-informed on AI will naturally lead to a more holistic strategy about how AI is to be used in collaboration with humans going forward. Finally, AI literacy and diversity in participation will ensure that the goals of AI align with ours as humans.
2. Bridge the “gap” between technologists and problem-owners to develop inclusive approaches.
The advancement of machine learning models, specifically deep learning, relies on the ingestion of data - structured or unstructured. The sharing of this data from problem-owners to technologists is the key to developing not only innovative but inclusive AI solutions. A better AI future built on diverse data sets requires these two parties work collaboratively. Projects like the AI Commons-- a sandbox created by AI technologists to get the data they need to test solutions -- are the first step towards bridging this gap and joining these two parties.
From a government/NGO-perspective this means continuing to provide open and structured data sets for AI practitioners to use; while still protecting the sensitive information that keeps our citizens safe -- a delicate balance. Providing these data sets is the first step but making others aware through promotion campaigns is a second critical step.
3. Harmonize the AI for Human Benefits Dialogue
To build a better AI future, as a society, we need to understand how AI can be used to supplement and benefit humanity. Because of the data revolution - we now have the data to build AI technology aimed at solving challenges that were previously unsolvable. This is one place where organizations participating in the AI community can add extraordinary value.
XPRIZE is a formidable example, it brings people together to participate, captures data, demonstrates its use, and stimulates technological advancement for the greater good. Through the incentivized competition model, XPRIZE challenges competing teams to demonstrate how AI can be used to solve a whole range of challenges - from solving biometric identity among children to improving global literacy through child-driven learning to improving emergency management responses and disaster resilience. This helps to provide concrete and practical proof points around how AI can be leveraged for a better future, while also helping to harmonize the dialogue around ethical, safe, and responsible AI.
While these actions and steps are grand and not without their unique challenges - they do help to begin to lay out a path for a brighter and better AI future amongst the race to AGI. Arriving at this future-state requires active buy-in and engagement from every single one of us - private citizens, NGOs, government organizations, and private companies. Each of these groups has a role to play in building a better AI future. The rate of progress is exponential, and so we must be vigilant and root our actions in purpose every step of the way, and we must act together, leverage our disagreements constructively, share our ideas, find consensus on goals, and continue to prepare for AGI.