Turbulence is widely recognized as the leading cause of injuries to flight attendants and passengers on commercial air carriers. Oceanic and international routes are subject to clear-air turbulence, mountain wave turbulence (which for Atlantic flights is particularly significant over Greenland), and convectively-induced turbulence. Turbulence encounters may occur in remote regions where ground-based observations are sparse, making hazard characterization more difficult, and where international turbulence and convective SIGMETs provide only low temporal and spatial resolution depictions of potential hazards. Therefore, a new effort is underway to develop a global diagnosis and forecast system that will augment and enhance international turbulence and convective SIGMETs and provide authoritative global turbulence data for the NextGen 4-D database. This fully automated system, modeled on the FAA's Graphical Turbulence Guidance (GTG) and GTG Nowcast systems, will employ NCEP Global Forecast System model output and data from NASA and other national and international satellite assets to produce quantitative turbulence nowcasts and forecasts. The convective nowcast methodology will make use of GFS data and operational satellite data from GOES, Meteosat and MTSAT, and will be tuned and verified using data from TRMM, Cloudsat and MODIS. Satellite-based turbulence diagnosis algorithms will also be developed. AIREPs and AMDAR data will be used in conjunction with a machine learning methodology to develop an empirical model that maps the model fields, turbulence diagnoses and convective nowcasts and derived features to global deterministic and probabilistic nowcasts and forecasts of turbulence. This paper describes the first steps toward using artificial intelligence techniques for data analysis and algorithm development.