miércoles, 21 de noviembre de 2018

AI and human rights

There´s a race among the different economic areas for a quick and effective deployment of Artificial Intelligence. The promise of a new surge of productivity growth accompany Artificial Intelligence is fueling investments in all the continents. The promise is an increase of 14% on global GNP by 2030, the private investments for reaping AI benefits are expected to reach $232B by 2025.

Nevertheless, besides its economic benefits, AI is carrying fears of a soul-less world where human beings will be subdued to machines. These fears that are captured human imagination would almost be an unavoidable future without a general consensus of the limits to be respected. A well positioned forum to hold this debate are the United Nations, where documents as the report on AI implications for human rights published by the UN Special Rapporteur for the promotion and protection of the right to freedom of opinion and expression are debated.

The report focuses particularly on the impact of AI on rights to freedom of opinion and expression, privacy and non-discrimination. In order to begin from a right starting point, the report includes a definition of AI as a “constellation” of processes and technologies enabling computers to complement or replace specific tasks otherwise performed by humans, such as making decisions and solving problems. Another cornerstone of the report is the human responsibilty as central in AI development. The author highlights the human role in in the design of AI systems (by the definition of its intentions), deployment (through the modellation of algorithms and the procurement of data) and implementation (in the decision of how to apply its outputs).

In order to be effective, the reporter narrow the scope of its task to concrete applications of AI. However, the selection does not mean a reduction of the relevancy of conclusions of the report, as the applications chosen are "Content display and personalization", "Content moderation and removal" and "Profiling, advertising and targeting". It´s not needed to say that the above mentioned applications are present in the daily surfing of internet users.

The reporter raises concerns in several areas related to human rights. Firstly, for its impact on right to freedom of opinion and expression, as AI endangers individuals’ self-determination and autonomy to form and develop personal opinions based on factual and varied information. Secondly, regarding the right to privacy the reporter warns on the mass collection of data that once processes may drive to sensitive information about people. Last but not least, the reporter points to the threat of non-discrimination that poses AI.

Governments are pointed by the author as the main responsible for pushing towards a human-centric development of AI. But he also identified several tools that could help to enable a right regulatory framework for AI:
  • human rights impact assessments performed prior, during and after the use of AI systems;
  • external audits and consultations with human rights organisations;
  • enabled individual choice thanks to notice and consent;
  • effective remedy processes to end human rights violations.

To sum up, AI development should not only fulfill the principle of ethic-by-design, there is also a human right framework to take care of  and tools for its enforcement that governments could use.

No hay comentarios:

Publicar un comentario

palyginti kainas