We recognize the challenging ethical aspects involved in the development of a society dominated by advances in AI. However, we also believe that this issue must be faced in a mature, open mind and responsible way. It is essential not only to respect the citizen privacy but also to maintain an alert attitude towards any possible security breach and other peril associated with the access to personal and public data. We believe that the increasing number of AI applications will have a profound impact far beyond anything we have imagined. This promising future must encourage us to face all possible fears associated with the development of AI.
We must build trust in society maintaining the highest standards of ethical behavior with a deep respect for human and civil rights. We must guarantee complete transparency of our operations and take into account the moral and ethical concerns of the society. Algorithms that will, to some degree, affect the citizens’ life must be accountable to those same citizens. The treatment of datasets, and particularly those containing personal data, must always meet the principles of data protection by design and by default.
The guiding directive of AI2 is that research and development of Artificial Intelligence must always serve humankind. The frightening possibility of an increase in unemployment, and its inevitable consequence of unbalance in the wealth distribution, the risk of impoverishment of interpersonal relations or the abuse or malicious use of this powerful technology can, and will, be addressed in due time. Nonetheless, the guiding directive translates to a set of conducts that the members of AI2 commit to following. For instance, over the next years, several social and regulatory frameworks and a constant assessment of the benefit-cost ratio of AI will have to be put in place, and AI2 commits to abide those regulations when they come to fruition. Until then, a lot of work must be done to establish the frontier of the AI reach and, at the same time, address any public mischaracterization of AI with facts and science-based arguments.