Limits that Artificial Intelligence has
Artificial Intelligence- A Cognitive Task
From Siri to self-driving cars, artificial intelligence (AI) is progressing rapidly. While science often portrays artificial intelligence as robots with human-like characteristics, they can surround anything from a smartphone’s voice commands to Google’s search algorithms. Although AI is considered a huge help in our society today, there are situations where they could be a disadvantage. Artificial intelligence contains risks and conflicts through its advancements on visual perception, speech recognition, and decision making because they only have a limited task, to do what their manufacturer made them to.
Artificial Intelligence was made like a human eye. Examples including Apple’s face ID, Samsung’s Iris Scanner, and Facebook Tags all have visual perception that are able to distinguish a certain type of face, therefore, giving access to the user. But these do not work the first time, the user has to give the command to let the machine know that they are being recognized, then, they input their data and observations into their system. AI’s visual perception works when shown the same or similar images and scenarios enough times. A facial recognition is a super advanced software that requires high quality digital cameras for algorithms to operate accurately. A face-detection system captures a face in the photo or a screenshot from a video, then the relative size of that face image will be compared with the size of the registered one. Therefore, the quality of the captured image alters the process of facial recognition, how well it would be done. Most systems allow recognition of a face-size range to reject invalid recognition and speed up the process of the image. To incorporate a face through the recognition software, multiple angles are being used. In order to generate a clear guide for the face, a frontal view is needed, much like most phones that include a face ID in order to be unlocked. The higher resolution the photo has and the more direct its angle is, the more accurate resulting matches would be. Then, there are also troubles with other things such as facial hair or sunglasses. An individual can still get away with facial recognition with a suddenly appeared or removed facial hair, same goes for obscuring face parts with glasses or masks. To avoid such failures, the databases must be regularly updated with the most up-to-date images. This is one of the problems that occurs regarding AI. It is not accurate to the user. It takes time to recognize what the user is showing, and if something similar is in sight, it may react the way that the user does not want to.
Although AI has a huge advancement on speech recognition, including Siri, Google, Alexa, and Cortana’s Voice command, they occasionally malfunction since they could only handle so much. “Speech is an attractive option when hands-free operation is necessary, as when operating machinery” (Russell 913). However, all of them would not work without an active internet connection. Most of the data they are processing relies on the data they find on the internet. Although the command sounds very simple, they all rely on cloud computing. Cloud computing is using a network of remote servers hosted on the internet to store, manage, and process data, rather than taking up space of the local server, and personal computer. They may stumble upon listening issues. Most of the time, Siri works just fine, but in order for Siri to understand the command and actually do what the user wants, they need to speak clearly. Siri might not comprehend a fast paced command, or a strong accent. They can also be literal at times. If they do not understand what the user is asking about, they will most likely tell the user to search it on the web themself. They do not understand spelling variations. It may successfully match common names like “Sarah” or “Mike” but may have trouble comprehending less common names, unless programmed it on the app by the user.
Decision making is the most impactful feature of AI to society today. According to an article on ProQuest titled “Artificial Intelligence and mobility- a proactive road to process control,” there is currently a shortage of time to collect data in factories due to its large amounts that help engineers keep their processes stable and in control. It also takes time to figure out what database from which to pull the data from, to get reports, and to analyze them. That’s why they want to leverage artificial intelligence to cut the engineers’ work in half, or more. To shorten time of the collection and analyzation of data. AI is providing automation that helps predict potential process problems, and goes to the root of the problem before production loss happen. But there are conflicts to these. These machines are programmed only to do one job. They are not responsible of duties outside their scope. They have to undergo maintenance, etc. Although AI is a huge help, at the end of the day, the engineers are in control of them. AI has a limited function on decision making, they only get to decide based on what is provided for them. Going back on relying to cloud computing, whatever is in store for them, they will give the user. Although there are benefits such as sentence prediction, auto correct, and word suggestion on keyboards, traffic times, recommendations on shopping or restaurants, and music playlists, they all based it on the user’s actions. Every information they have are all catered on the individual’s usage. The most significant challenge that AI faces is the lack of security and privacy. Mobile apps that use machine learning techniques can collect user information, such as the user’s IP address and location. An employee that uses this kind of app for work might not feel comfortable using it on a personal device. Some companies distribute wearable devices to their employees that use AI technology to collect data for insurance reasons. These scenarios pose serious questions about employee privacy.
According to Irving John Good, a British Mathematician, “designing smarter AI systems is itself a cognitive task.” AI, no matter how intelligent it can be, is unlikely to exhibit human emotions like love or hate, and that there is no reason to expect it to become intentionally benevolent or malevolent.
- “Benefits & Risks of Artificial Intelligence.” Future of Life Institute. Jolene Creighton Https://futureoflife.org/wp-content/uploads/2015/10/FLI_logo-1.png, 2016. Web.
- Manfredi, Don. “Artificial Intelligence and Mobility-a Proactive Road to Process Control.” Plant Engineering 70.7 (2016): 45-7. ProQuest. Web. 27 Nov. 2018.
- Russell, Stuart J., and Peter Norvig. Artificial Intelligence a Modern Approach. Boston: Pearson, 2010. Print.
- Solutions, TecSynt. “Pros and Cons of Facial Recognition Technology For Your Business.” Hiring | Upwork. N.p., 25 Sept. 2018. Web. 01 Dec. 2018.
- Vincent, James. “These Are Three of the Biggest Problems Facing Today’s AI.” The Verge. The Verge, 10 Oct. 2016. Web. 01 Dec. 2018.