A decade of high-resolution monitoring has revealed extensive activity in fresh Martian gullies. Flows within the gullies are diverse: they can be relatively light, neutral, or dark, colorful or bland, and range from superficial deposits to 10-meter-scale topographic changes. We observed erosion and transport of material within gullies, new terraces, freshly eroded channel segments, migrating sinuous curves, channel abandonment, and lobate deposits. We also observed early stages of gully initiation, demonstrating that these processes are not merely modifying pre-existing landforms. The timing of activity closely correlates with the presence of seasonal CO2 frost, so the current changes must be part of ongoing gully formation that is driven largely by its presence. We suggest that the cumulative effect of many flows erodes alcoves and channels and builds lobate aprons, with no involvement of liquid water. Instead, flows may be fluidized by sublimation of entrained CO2 ice or other mechanisms. The frequent activity has likely erased any features dating from high-obliquity periods, so fresh gully geomorphology at middle and high latitudes is not evidence for past liquid water. CO2 ice-driven processes may have been important throughout Martian geologic history, and their deposits could exist in the rock record, perhaps resembling debris-flow sediments. Gully landforms on Mars resemble water-formed features on Earth, with channels transporting material from an alcove to a depositional apron. From their discovery (Malin and Edgett, 2000), they have generally been interpreted as evidence for wet debris flows or flowing liquid water (e.g., Carr, 2006). Such liquid would have major implications for Martian climate, geology, the possibility of life, and the definition of Special Regions for planetary protection (Rummel et al., 2014). Understanding the formation of gullies has thus been a major focus of recent Mars science, as shown by the work in this volume. Numerous hypotheses for gully formation have been considered. Martian surface conditions are not favorable for the existence of liquid water, so initial models focused on release of groundwater from shallow or deep aquifers (Malin and Edgett, 2000; Mellon and Phillips, 2001; Gaidos, 2001), possibly aided by geothermal heating melting permafrost (Hartmann, 2001; Hartmann et al., 2003) or the occurrence of brines (Knauth and Burt, 2002). However, the occurrence of gullies on sand dunes and isolated peaks argued against significant input from aquifers, and led to the development of models based on insolation-driven melting of snow or shallow permafrost at times when the