Updated: Sep 25
Source: MIT Technology Review
Determined to build more powerful algorithms, researchers are heavily reliant on the usage of computing power, centralized cloud services, and greater amounts of data. However, this intense usage has multiple negative outcomes and to overcome these problems, tech giants and qualified researchers are currently developing newer algorithms “to shrink existing deep-learning models without losing their capabilities.” In the meantime, a new generation of AI (Tiny AI) assures to meet the goals of these researchers! This new version of AI holds more computational power into a smaller space and runs on less energy making it much more efficient. Some of the advantages to Tiny AI include better and faster voice assistants, digital cameras, and autocorrect without needing to alert the cloud whenever these services need to access “a deep learning model.” Moreover, Tiny AI ensures to protect our privacy in a more secure way and allows for newer applications to be made. Nonetheless, with its benefits come its challenges. Combatting “surveillance systems or deep fake videos, for example, and discriminatory algorithms could also proliferate.” Fortunately, with the work of engineers and policymakers, a new fixation could be made to check on these potential harms.