Autonomous vehicles like Waymo(a self driving service) rely on machine learning and computer vision for perception and decision-making. They use deep neural networks to detect and classify objects (vehicles, pedestrians, signs) from camera and sensor data. They fuse inputs from cameras, LiDAR, radar, and ultrasonic sensors to build a 360° model of the environment. To recognize this data they use aggregate data of other examples to get as close to 100% recognition. Advanced sensor  and reinforcement learning algorithms enable path planning and control over route planning. Computers with extremely powerful cpus and gpus run these AI models in real time (Autonomous Vehicles: Evolution of Artificial Intelligence and the Current Industry Landscape). The benefits are improved road safety due to improved reaction speed over people, but that is also its biggest challenge(The evolving safety and policy challenges of self-driving cars). They have to perform complex ethical challenges like a human would. 

Insights: Self driving automobiles have the possibility to take over driving as we know it, but we have to make sure we are doing it at the right pace. If we do it to fast it could perform fatal errors, unnecessarily wounding people and lowering people's perceptiveness towards it.  We also have to make sure that we have adequate jobs in other sectors for people who will lose their jobs. However if we successfully implement it, it will be a net positive as cars on the road can communicate with each other to prevent accidents.

We see traffic lights and signals everywhere, especially in dense places like New York City, but they can often be inefficient and lead to slowdowns. With AI this has a possibility of being greatly reduced. Cities like san francisco and los angeles are using AI from streetlight data that aggregates cell phone usage to create an optimal flow of traffic and predict congestion patterns according to tractiontechnology.com. They can automatically adjust traffic light cameras so instead of a fixed distance it uses optimization algorithms, coordinating with nearby lights. According to smartcitiesdive.com this cut idling time by 40% and traveling times by 25% at equipped stations in pittsburgh. However this creates privacy issues where many drivers don’t want their cars to be constantly recorded and stored for police reasons.

In my opinion this could be an amazing way to clear up congestion in major cities because traffic wastes a large portion of people’s lives. I think more research should be done on how to make the transition smoother between our current system and a more modern one, including having two systems run simultaneously to prevent shutdowns.

In logistics AI can be used for predictive analytics, route optimization, and automation. One major use is predicting when a product will be needed by a consumer so it is already close to them. Amazon does this particularly well netting the one day delivery times. Logistics uses machine learning models to analyze a lot of factors, including historical sales, weather, and market data to predict product demand more accurately than manual methods. They also use it for the trucks, predicting when it needs maintenance or optimizing their routes according to AI In Logistics: The Trends And Benefits Transforming Supply Chains).  UPS’s ORION system uses Ai analytics to map out the most efficient daily routes for tens of thousands of delivery drivers which saved millions of miles according to (Route Smarter, Not Harder: UPS's ORION Success). There are very few problems with this when executed successfully but in unforeseen circumstances like routes being shut down you still need human input.

In my opinion this could be very useful, especially relating to the environment. If shipping routes are cut down upon it could save a large amount of gas, helping reduce pollution. One thing that potentially worries me is it makes buying things online to easy, and can make people lazier.

Studying ethics and safety in humans is hard enough, but AI presents new challenges. One major issue is deciding who is responsible if there is an accident on the road, with regulators are struggling to create liability frameworks to ensure accountability. One of the most important factors in this is transparency. If people know more about how the AI works and what affects their decision making, people can make the choice for themselves. The problem arises when AI is used without peoples consent, like when a AI dash cam recorded a trucker and its location data (AI dashcams enhance trucker safety while raising privacy concerns | Fox News). AI must also consider bias and equity issues. If it is trained on data that is biased toward wealthy people and neighborhoods, it could prioritize traffic flow there at the expense of the poorer one. (AI in Transportation: Ethical Balance between Safety and Privacy). 

I think it is extremely important that we look at AI with a careful lens to see if one group is getting unfair treatment. I remember in middle school as a project we had to google general topics and see what types of people were represented. It was usually white, wealthy, heterosexual people and if AI is trained on this data it could prioritize them more.

AI taxis or AI-enhanced mobility services could change people’s lives but it could just prop up an already unfair gap between the rich and the poor if companies only target wealthy cities. Planners are already working to create affordable transportation options for with AI, such as subsidized autonomous shuttles in underserved areas (Equity, Artificial Intelligence, and Transportation). We also must make sure that bias in algorithms is removed. If an AI for traffic management is trained on data that reflects a police bias towards a specific race or class, it could further exacerbate the inequality we have in the US.

• UPS. (2024, March 14). UPS strengthens network capabilities and employee experience through innovation. [Press release]. UPS. https://about.ups.com/us/en/newsroom/press-releases/innovation-driven/ups-strengthens-network-capabilities-and-employee-experience--th.html
  
  

• TED. (2023, June 5). How AI will revolutionize the way we live and work | Kai-Fu Lee [Video]. YouTube. https://www.youtube.com/watch?v=-_wNjLduGiY
  
  

• Matellio. (2023, October 11). AI traffic management: Revolutionizing smart transportation with artificial intelligence. Matellio. https://www.matellio.com/blog/ai-traffic-management/
  
  

• Freethink. (2023, December 15). How AI is changing transportation [Video]. YouTube. https://youtu.be/tgYOOPMQmc4?feature=shared
  
  

• AI Business. (2024, February 8). Waymo launches AI model for autonomous driving. AI Business. https://aibusiness.com/nlp/waymo-launches-ai-model-for-autonomous-driving
  
  

• Veritasium. (2023, November 22). The future of self-driving cars [Video]. YouTube. https://youtu.be/taMP_n3wL7M?si=aFMC--utl_AYwLIC

• iRideTheHarlemLine. (2021, February 18). Behind the scenes at the subway’s power control center. iRideTheHarlemLine.https://www.iridetheharlemline.com/2021/02/18/behind-the-scenes-at-the-subways-power-control-center/