Project Maven: Ethics concerns from Google employees

Project Maven: Is AI drones ethical or not?

Link: https://gizmodo.com/google-plans-not-to-renew-its-contract-for-project-mave-1826488620

Summary:

In 2017, Google was awarded a defense contract for Project Maven. The entire purpose of Project Maven was combining machine learning and engineering for the US drone system to distinguish friend from foe. The ethics concerns that some Google employees have with this project is that it can scan and memorize the faces of individuals from a distance using footage from the internet. With the power and the information on the internet on a global scale and being able to determine whether or not this individual is a friend and foe can be great for military applications. However, the thought of killing individuals who have been determined to be a foe for military personnel is concerning these individuals. Because of this reason, more than 10 Google employees left with others threatening to leave the company. To prevent this, Google said they were not going to further pursue Project Maven when the contract ended in March 2019.


Jason’s POV:
The question of ethics is always something difficult. After being in the shoes of servicemembers in combat, I can say that being able to ID friendlies and foes in seconds before engaging can mean a lot. Not being able to quickly determine who is friendly and not can mean life and death. Being able to help the warfighters with accurate information I believe is ethical. If it means civilian and friendly casualties dropping will be the greatest thing for any military group. I always had the thought that isn’t saving innocent lives ethical? War is a constant issue that plagues the world today. It’s not going to disappear anytime soon as there is constant conflict around the world. But if you can reduce the chances of death of a civilian would you not try to attempt this? Which leads me to another point, if employees felt the need to be ethical why not try to use the position of Google to try enacting policies to ensure it doesn’t spill over to anything else if that is the worry? Rather than fixing the issue, they run away from the issue.

Alice’s POV:
We are in a technology phase where it's hard to differentiate what is ethical or not. Is having an artificial intelligence program incorporated in a drone that would help surveillance the area for danger, ethical or not ethical? I think the identification of friend or foe would be great for the military to have to protect the citizens, but I can understand why it’s a ethical concern as Google operates in multiple continents, they house employees from all over the world, so if it’s used in US military, I can see why some employees are concerned about it. Especially when it means that Google AI team are in a way apart of “killing” foes. But I feel like Google should have properly discussed the pro and cons, and really think through before signing the contract to be apart of Project Maven, as their pull out from it caused all the hard work, funding, and everything to be --in a way -- a waste of time.

Now the question for you is, what are your thoughts on Project Maven and how would you have done to deal with the ethical concerns that are rising as we are trying to come up with better ways to protect ourselves from outside threat.
             

Other links on Project Maven for reference:

https://globalnews.ca/news/4125382/google-pentagon-ai-project-maven/

https://globalnews.ca/news/4124514/google-project-maven-open-letter-pentagon/

https://theintercept.com/2019/02/04/google-ai-project-maven-figure-eight/

Image Recognition Technology

Image Recognition Technolgy in real life

The amount of information on the internet is so huge. For companies, this huge amount of information is both challenge and opportunity to manage effectively market themselves, protect their image and utilize them in the era of flooding information. Image recognition tools are an important key to the potential of huge information.

What is the image recognition technology?

It is simply that the computer uses your eyes as its own “eyes”.  Image recognition is the software ability that processes all pixels that makeup images. So, the network provides a vast number of images to the computer, and computers are able to learn and recognize objects. In this process, the software for image recognition has to be required deep machine learning.

For instance, computers would be shown millions of images of apples. And then, the computers learn what images of apples are, and how looks like an image of an apple. If computers see the image of the banana, they would compare all the pixels of the image of the banana to the image of an apple. And then, if the computer cannot find a similar pixel of the image or the image does not match nearly, the computer would recognize the object as the banana.

Why do we should use image recognition?

Image recognition tools can be used in various businesses. By using image recognition tools, companies can know the interest of customers. Using this data, a brand can do targeted marketing for specific customers. For example, Social networking service provides personal preference advertisement to the people. For the people who love shoes, SNS provides advertisement related to the image of shoes.

In the business industry, many image recognition tools are being used.

• Price comparison: People can take a photo of the product that they are interested in buying, and then they can find the near store to buy it by using the application like Google Shopper.

• Driverless car: Using the vehicle’s visual system and image recognition, driverless car identifies pedestrians, road sign, map, and movement of other vehicles.

• Image search: When people present Google with an image or URL, the Google search engine will show you the web site containing the image or same image that people presented.

The most innovative technology that used image recognition tool in daliy life is Driverless car industry

Drivers with the human can identify pedestrians, road sign or other vehicles very easily. However, it is not easy for the driverless car to do it. So, what are some of the ways that driverless car can identify and recognize the object? They are Edge detection and Image Classification.

• Edge detection

The most fundamental technique for image recognition is called Edge detection. The basic idea of this technique is that pixel related to individual objects will be similar, but the pixel that related to different objects will be different. Driverless car technology can use this technique to figure out the outline of the car and to track the movement of other cars

• Image classification

Image classification includes more advanced image recognition algorithms. This algorithm can be used in looking for what street sign it is, general shape, colors, and size of the sign. The common implementation of this technique is a neural network. For example, the computer might ask that “Is there a large red color in the image?” to itself, and then it might ask again like “Is it shape as an octagon or another shape?”. And then it might ask like “Does large red image and octagon object mean STOP?”. In this process, algorithm concludes that that object is the STOP sign

Tesla is the company that utilizes image recognition to revolutionize self-driving. It is called Autopilot technique.

Image recognition is available to all businesses without limitation in market size, or product. Image recognition can save your time and money. It can classify countless images and return the data that you really need. Image recognition is the key technology for solving complicated information in the real world.


Question

1. Have you ever experienced image recognition technology in your daily life? Explain your experience

2. Do you have complete confidence in Driverless car system? Explain the reason for your answer

3.  Do you have another innovative idea about image recognition technology? Please share your idea :)

Insecure Infrastructure & the Global Reliance on Outdated Systems

A lot of technology which the modern world has come to ignore or take for granted was developed very early in the age of computing and hasn’t changed in the decades since. For example, most financial institutions in the United States still use 1950’s-vintage COBOL code with roughly 75% of a single day’s business transactions executed via the language. You can still buy 5.250” and 3.500” floppy disks off Amazon for an insanely high price per megabyte, simply because there still exists specialized machinery that can only read from a floppy. Likewise, an obscure and low-level architecture born from the 1970s still drives a truly massive amount of physical systems.

Known as SCADA, or Supervisory Control and Data Acquisition, it exists largely unseen throughout industry and infrastructure. To the public, a SCADA system might only betray its presence via a small, unassuming antenna connected to a patchwork of grey metal boxes. 

Nothing to see here...

However, the scope of what is connected by SCADA includes air, railway, and automotive traffic control; water treatment; energy generation, transmission, and distribution; even automated manufacturing systems rely on it.

Historically, SCADA has gotten away with relying on “security of obscurity” – no one really knew about it, and for those that did, to actually exploit SCADA meant an understanding of a number of specific protocols. In recent years, however, this reliance has started to become a glaring vulnerability. Because of how specialized these industrial protocols are, they are not yet monitored by security systems. Many of the human-machine interfaces (HMIs) that utilize SCADA are physically insecure and theoretically can be accessed by anyone. On top of it, if there is any security present at all in the system, more often than not it is no more than the factory default login, which can be readily found in plain text lists.  See the list for yourself here.

So what are the implications of this? A Trend Micro report demonstrated how simple it was to shut down water flow at an industrial water heating center; additionally, they were able to shut off and reset oil rigs (which were easily geolocated using Google Maps); and they were even able to infiltrate a power plant, with the controls readily available to stop power supply. All these vulnerable sites were left wide open to the Internet.

In a different article, a SCADA pen-test was conducted on the operational technology network of an unnamed major airport. With no effort whatsoever, the test yielded something straight out of Die Hard 2: control over the jet bridge, runway lights – ultimately reaching total facilities control to the point of shutting down the airport.

Even railway systems are threatened, with computer based traffic control and switching systems readily attacked without needing much skill. Like the situation with the airport, it sounds like something out of Hollywood – per the linked article, it’s possible for attackers to be able to manipulate signaling systems in order to direct traffic onto blocked sections of track or even to engage a rail switch while a train passes over it (causing derailment). Although a bitter consolation, it should be noted that for an attacker to cause damage beyond simple mischief in the way outlined above does require ample skills. 

Or not, if your victims still run XP.

- - - - - -

Questions:

1) The last time the United States had a massive regional blackout (coincidentally, caused by computer issues) was in 2003.  Imagine that happening today.  What do you think the consequences would be?  

2) If/when smart grids eventually overhaul these legacy systems, do you think they will cause more problems than they solve?

Our Environment and Apple

This article talks about what Apple and other Silicon Valley businesses are doing to fight climate change. Currently President Obama’s former Environmental Protection Agency administrator, Lisa Jackson, works for Apple as their Vice president of Environment, Policy and Social Initiatives. Because of her work in Apple they have announced that 21 companies in Apple’s Supply chain will get all their electricity from renewable sources. According to the Washington Post article the this increase will bring the total amount of renewable energy used by the company to about 5 gigawatts. For comparable size 1 gigawatt of energy is equal to the power consumption of 7 Boeing 747 passenger planes or about 700,000 homes for a year. For other interesting power conversions checkout the Official Blog of the Navy.This also means that 40% of Apples energy consumption would be from renewable energy. Apple has achieved this by pressuring their suppliers to go green while also providing them with the tools to make the change, like the platform created by Apple to help suppliers find renewable energy sources to draw power from.

Lisa Jackson also stated that the majority of the suppliers that Apple got to switch to renewable energy where based in China. This is important because it shows other international companies that they should look at their supply chain and make sure that they are environmentally conscious even if they are located in a foreign country that has lax environmental laws. Even though Apple has shown the world they are trying to help the climate, many critics still point to the amount of electronic waste that Apple creates each year. Phone chargers and charging ports or “bricks” create a substantial amount of waste each year. I think that the critics may have a point because Apple creates new products each year and frequently makes older products obsolete. Lisa Jackson said that Apple was working on their charger problems which is probably on of the larger areas of waste that they create. But she stated that Apple had more pressing priorities like getting all the components of iPhones, MacBooks and other products made 100% from recycled materials. She stressed that this causes the most adverse effects to the climate because of the aluminum mining and smelting which creates excessive amounts of carbon emissions.

Today, many companies are taking an active role in combating climate change. This article says that corporations have begun to increase their environmental efforts after Donald Trump left the Paris Accords. For example, right now Google is working with leaders to identify large sources of Carbon pollution. Other Silicon Valley companies like Lyft have also started to create programs that are more environmentally and ethically conscious. Climate change is a problem that we will all soon face and, it is comforting to know that other companies are being socially responsible and are actively trying to reduce their carbon footprint without government intervention.

Questions:                                                                                                                                     1. Do you think that Apple should be criticized for the amount of “e-waste” they create each year, or should other environmentally conscious actions minimize the criticisms they receive?                        2. Should companies be responsible for setting their own environmental policies or should the government set the policies ensure that each company lessen their carbon footprint?

Why Marketers Should be Paying Attention to Twitch

An article posted about two years ago still remains relevant today regarding the internet streaming service Twitch. Twitch.tv is a live-streaming platform for video games, chats, and competitions for eSport genre events. Amazon bought the company in 2014 at the time for a bit less than $ 1 billion, but this to me demonstrates the significance Twitch carries as an industry titan like Amazon saw the potential in the platform's services. Unlike competitors such as Youtube or Netflix, Twitch has become something that is more than just video-streams, it has created a novel environment that integrates social interaction alongside a viewing experience.

The main point of this article is, why should marketers care about this seemingly small subgenre of video-gaming and streaming? Many people are unaware of the grossing profits and numbers this platform has been generating on average.

Some facts include:

• Over 100 million monthly viewers
• On average viewers spend more than 1.5 hours per day.

At this point Twitch is becoming more than just a "video game space" it has become a direct competitor to media outlets such as Netflix, iTunes, and Spotify. The article I am referencing which can be found here, shows that in terms of Twitter mentions, Twitch is keeping an equal pace with Netflix.

Twitch is becoming a large part of this generation's digital identity. One tactic going forward will be to see companies begin to create their own content, whether it is cooking, video games, arts and crafts or discussion, companies have plenty of opportunities to advertise within this space. If companies feel that Twitch demographics match the companies target consumers, it would be beneficial to start advertising. I personally just a week ago was browsing the Twitch.tv website and found Wendys, popular fast-food burger chain had created a profile and was live-streaming video games with over 8,000 viewers. During this viewing session, the viewer-chat was generally surprised and excited to see that "the Wendys" was partaking in their culture. Wendy's was showing live footage of someone playing Rocket League, a popular computer game, while having popup ads for burgers appear every now and then. Twitch may be unfamiliar ground for companies to get their feet wet with in regards to marketing, but I feel both this article and in my opinion, there is a lot of potential for companies to begin streaming and advertising their content in front of millions of viewers.

Link to Article: https://marketingland.com/marketers-paying-attention-twitch-202984

Q&A:

1) Can you think of some other examples of companies that could benefit from marketing via Twitch? If so who?

2) The most popular Twitch streamer has been able to generate $500k per month through the platform. Is this a good indication that Twitch streaming may become a full-time job in the future for some dedicated content creators? Is this a good or bad thing if you believe it becomes a career option.

Big Mac or Big Data?

In late March, fast food giant Mcdonald's announced their plans on acquiring an Israeli based tech company, Dynamic Yield Ltd. The company is known for offering retailers with "decision logic" technology. The offer is estimated to be over $300 million, making it the largest purchase by the company since Boston Market in 1999.

Over the past several years, you may have noticed many technological changes within many Mcdonald's locations, mainly with its menu displays. These technological investments have been made since CEO Steve Easterbrook took over in 2015. Mcdonald's plans on implementing this new AI predictive technology at the drive-thru and ordering kiosks inside the restaurant. In an exclusive interview with Wired, Easterbrook said, "How do you transition from mass marketing to mass personalization? To do that, you’ve really got to unlock the data within that ecosystem in a way that’s useful to a customer.”

In a pilot program in Miami, the algorithm used data not only from historical orders at the location, but also local time, date, weather, traffic, and even nearby events. It can track trends at the location and around the globe causing more potentially effective upsells.

Personalization will also be focus with the adaptation of the new technology. They plan on using user information (if they are willing to provide) to create a "personal" menu when they arrive. Mcdonald's already uses geofencing technology to know when mobile app customers is approaching to begin the order. This technology can raise privacy and security concerns, so it will be interesting to see how Mcdonald's goes about protecting it. “We will be very sensitive as we learn, as we go forward,” says Easterbrook. “I think over time it’s going to be important to demonstrate that we can offer value back for customers willing to open themselves up to us.”

Questions:

1. Do you think AI predictive technology will affect the way you order at the drive-thru?

2. How soon do you think it will be until other major franchises begin adopting this technology?

3. What privacy concerns, if any, can you think of with this technology being implemented?

Weaponization of the Internet of Things

by Alex Windle and Maddie Stidham

What are IOT Devices?

The Internet of Things (IoT) refers to something that most of us interact with every day, but may not realize. IoT is the network of everyday, physical objects that are equipped with Internet connectivity. The main goal of IoT devices is to collect and share data about the way they are used and the environment in which they function. They can be embedded with electronics, Internet connectivity, and other forms of hardware and can be remotely controlled and monitored. From smart refrigerators, to self-driving cars, to our fitness watches, the valuable information that these devices can track opens us up to some serious security and privacy concerns. Their pervasive presence is supposed to make life easier, but there is a darker side to these new emerging technologies--the weaponization of the Internet of Things.

Botnet and DDoS?

A Botnet is a group of IOT and traditional computing devices that have IP addresses and have been compromised. A botnet could be just two devices, or as much as Mirai; 600,000. A botnet then can be used in a DDoS or Distributed Denial-of-Service attack to overwhelm a system with a massive amount of data. This usually happens to web servers but can even bring down connectivity to large geographical areas like Mirai taking down US east coast connectivity. Essentially, a botnet performing a DDoS attack is like a bunch of zombies creating a road block between you and Illegal Pete’s on a Thursday night.

Organizations and people create botnets to perform DDoS attacks for a multitude of reasons. Companies perform attacks in order to shutdown competitor's E-commerce website. People can use DDoS attacks to keep a small or medium sized company’s website down, then demanding money to stop their attack. This is similar to the mafia ‘requesting’ money to ‘protect’ you. Another example of DDoS attacks is humorously, competitive online games to block their competitor's packets to the game's server so that they can gain a competitive advantage. There are many more examples here.

Can I make my devices secure?

As more and more of these devices make their way into our homes, cars, and businesses, the rush for convenience might leave security as a non-priority. Considering the serious security concerns surrounding Internet of Things devices, there are steps we can take to secure those devices. The following are our team’s top three tips for securing your Internet of Thing devices:

Make sure you have the latest firmware

Major companies prepare for security threats and patch the vulnerabilities that they become aware of. That being said, one must update their devices in order to receive those security updates, they are not just trying to stop your 12th binge of The Office. It is important to remember that there are lags between when vulnerabilities are discovered, and when they are patched. See the following to remain more secure during those times.

Disable Universal Plug and Play (UPnP)

Universal Plug and Play is a set of protocols and related technologies that allow devices to automatically discover one another. This feature, though convenient, serves as a gateway for hackers to infiltrate your IoT devices and network. According to Dan Goodin, “security experts are advising that a networking feature known as Universal Plug and Play be disabled on routers, printers, and cameras, after finding it makes tens of millions of Internet-connected devices vulnerable to serious attack”. If you want to avoid cyber criminals potentially discovering your devices from beyond your local network, simply disable UPnP completely.

Pick strong, unique passwords for every device

Strong passwords are not necessarily passwords with 6 characters, 1 uppercase, and 1 special character. Strong passwords or pass-pharases are generally supposed to be fairly large (12 characters or more), since brute-force password crackers take exponentially longer to hack with each additional character. Password managers such as LastPass are helpful for managing and creating large randomized passwords if you need help. Here is LastPass’s free password generator.

Questions

1. Have you heard of IoT devices being used for cyber attacks before reading this post? Is so, in what context?
2. How many IoT devices do you own and have you thought about securing them before?
3. Do you use a password manager? If so, which one and why?