Artificial intelligence has been increasingly transforming the industry, the financial industry is one of them. various advanced technology has been applying in banking and financial service institution. The major segments that banking and financial service institutions apply machine learning into our decision making, tailor service and risk management (McKinsey & Company, 2019). Research shows that the use of machine learning brings more than $250 billion in revenue into the banking industry (McKinsey & Company, 2019). previously, the banking data analyst uses a vast of data to manage risk by selecting model, cluster data and analyzing data. However, it is no longer the case with the introduction of machine learning. The algorithms can do these routine jobs much more efficiently. Expert in the field of ai says that machine learning algorithms are able to outperform human in the most situation(Nawijn, 2019). However, does it mean that machine learning is perfect and human interaction in banking risk management is not necessary anymore? I would say no.
Although machine learning could solve most of the problem and have superior performance than human, there are few potential risks and flaws exist. for example, model risk. Model risk occurs when a financial model is used to measure quantitative information, wrong input or the inadequate model could lead to model risk thus results in wrong output (Investopedia, 2015). Conscious of the model risk, many banks proceed cautiously and restrict the use of machine-learning models to low-risk applications, additionally, the bank should employee supervisors or motors to regularly exam the work of machine learning and validate the results. Therefore, the combination of machine learning and traditional framework associated with risk management is crucial for banking.

Besides risk management, banking also employe machine learning to make an investment decision and manage the portfolio. The algorithm enables the bank to generate a more rational decision based on the thousands times of simulation and calculation. However, since the systematic risk is unpredictable and could not be eliminated, it is very subjective to make the decision that relates to the overall market situation. therefore, we still need human interaction in the decision-making process.

-McKinsey & Company. (2019). Derisking machine learning and artificial intelligence. Available at: https://www.mckinsey.com/business-functions/risk/our-insights/derisking-machine-learning-and-artificial-intelligence

-Nawijn, B. (2019). Managing financial risk with Machine Learning. Tjip.com. Available at: https://www.tjip.com/en/publications/managing-financial-risk-with-machine-learning .

-Investopedia. (2015). When Model Risk Occurs in Finance.  Available at: https://www.investopedia.com/terms/m/modelrisk.asp.

The last mile carries, the last but not the least phase of delivery, determines the quality of express service. Timing and accurate location make up the touchstone of success. However, people seem not satisfied with the current service any longer, they are eager something even faster, one-day shipping.Before we dive into one-day shipping, let’s talk more about last-mile delivery first.The last mile, which is the last step of delivery service, it makes up the movement of your parcel from the hub (distribution center) to your final destination. It might sound easy, however, lots of last-mile delivery service seem as inefficiency. So what is the reason behind the problem?

Cost, which is the most significant issue along with the last mile. It is surprisingly expensive to deliver different, unique parcels and shipments to the specific, unreliable destination on any random spot on a delivery route (Connecticut Courier Service, Expressway Courier, 2018).Besides the heavy cost, unpredictability is also a serious issue. Often time, the carriers have to deal with the situations that customer is not home, late delivery due to bad weather or traffic, etc. Those unpredictable events make the last-mile delivery even costly and rigorously.

Amazon Flex, similar with uber, which allows people to be an independent delivery partner and earn money in retuen based on the on-demand contract (Schoolov, 2019).
Anyone over 21 with a driver’s license, car insurance could be a member of Amazon Flex. The independent carrier could gain around $18- $25 per hour in return. To smooth the delivery service and appeal to more people to join in Amazon Flex, Amazon develops an APP for this program. The user could execute a private delivery service via the app. From scheduling a delivery time, picking up the parcel to delivering the parcel. All the process could be done easily online.

This new business model takes advantage of digital devices and data to break the vast amount of workload into numerous small and individual tasks. On the one hand, it dramatically decrease the delivery cost, on the other hand, it makes one-day shipping come true. We can detect the threat arise from this new business model if the independent delivery service could cover more regions. Therefore, the market share of traditional logistics company might be shrunk unwillingly and the the war between the big logistics company and  the private contractor is inevitable.

However, there are few potential risk and challenges cannot be ignored.

1. Labor legitimation
2. Supervision and control
3. After-delivery service
4. Parcel stealing 

Matching each delivery service with the person who executes this action should be implementation accurately. Amazon flex should be able to track individual carrier and their car in real-time, just in case parcel stealing.

reference：

-Connecticut Courier Service, Expressway Courier. (2018). What Is Last Mile Delivery and How Is It Changing?. [online] Available at: https://expresswaycourier.com/what-is-last-mile-delivery-and-how-is-it-changing/ [Accessed 3 Oct. 2019].

-Schoolov, K. (2019). What it’s really like to be an Amazon Flex delivery driver as Prime one-day shipping expands. [online] CNBC. Available at: https://www.cnbc.com/2019/06/19/how-amazon-flex-delivery-drivers-get-paid-and-what-its-really-like.html [Accessed 3 Oct. 2019].