CS 540: Stanford AI Report Essay

AI as a field has had its up and downs since the inception of Computer Science. Today, one could argue we are approaching an AI renaissance, where most basic tasks will be automated, and the field will be integrated with several existing fields such as finance, medicine, education, transportation etc. However, this naturally presents some challenges, which this short essay plans to discuss, in regards to the Stanford One Hundred Year Study on AI report.

The paper argues that robot cleaners are restricted to a limited region, usually a house or a room, and mechanical problems such as single step stairs, and other mobility related issues. These issues, while relevant a few years ago, can be considered trivial today. Great developments in robotics by companies like the iRobot, Boston Dynamics etc, have shown compact robots easily dealing with such issues with little to no difficulty. Another issue the paper discusses is about getting perceptual algorithms and 3D Object recognition and mapping to work on small devices such as the Roomba. As the case of Snapchat proves, various Computer Vision algorithms can be made to work for a relatively small device such as a smartphone, while still yielding efficient results. By the end of this decade, with a decent increase in memory, these algorithms, paired with the right sensors, can easily fit on to these small devices, thereby making them more efficient, and better at their tasks.

The report also discusses about how home automation can become a huge part of peoples lives, as we have already seen with devices such as the Alexa products, and the Google Home. While these devices are certainly a sign of a more automated future within households, it still holds the significant risk of lack of adequate security measures. Recently, articles have surfaced about Echo Dot owners hacking their devices and gaining root access to them. This enabled them to receive every piece of audio the device captures, even if it isnt a direct command, therefore becoming a live mic in the wrong hands. While the procedure requires a few hours, it still holds a threat in the form of second hand devices, or a device left hidden in a common room. While Amazon has reportedly fixed this bug, it still raises many questions about how secure and safe these gadgets exactly are to the common user?

While the concept of AI and robots interacting with humans isnt exactly new, thanks to movies like 2001: A Space Odyssey, Terminator, Blade Runner etc, they are still breaking new grounds in fields like medicine and education. While the concept is exciting and fresh, even after many theories and ideas, it still raises the questions of whether the technology has caught up with our imagination. Especially in the field of education, while most tasks, such as standardized test checking, have become automated, we still have a long way to go. While these devices are helping teachers teach and manage, they have yet to prove useful to teach advanced topics to students. For example, if a student were to have a specific problem, a teacher could ask questions to understand their exact confusion, and then remove it, and try to solidify the students understanding. AI and NLP have not reached this level yet to completely understand a students doubts and then clear them. This would especially prove useful for young children who find difficulty in identifying their problems concretely.

While the recent developments in AI have been fantastic, much more can be expected from the future generations of AI. It will indeed be exciting to see how the technology develops to further help humans with their day to day tasks.
