Lecture
AI systems are applied in various fields
Financial institutions have long used neural networks to detect suspicious events and actions.[41] The use of AI in banking began as early as 1987, when Security Pacific National Bank in the United States created a task force to combat fraud and unauthorized use of debit cards.[42]
Algorithmic trading involves the use of sophisticated artificial intelligence systems to make trading decisions at a speed exceeding what the human body is capable of. This makes it possible to execute millions of trades per day without any human intervention. Automated trading systems are typically used by large institutional investors[43].
Several major financial institutions have invested in developing AI to use in their investment practices. BlackRock's AI system, Aladdin, is used both internally and by the company's clients to assist in making investment decisions. The system's wide range of functionality includes natural language processing to read text such as news, broker reports, and social media feeds. It then assesses sentiment about the companies mentioned and assigns them a score. Banks such as UBS and Deutsche Bank use an AI system called Sqreem (Sequential Quantum Reduction and Extraction Model), which can process data to build consumer profiles and match them with products they are most likely to want.[44] Goldman Sachs uses Kensho, a market analytics platform that combines statistical computing with big data and natural language processing. Its machine learning systems draw on data from the Internet and assess correlations between world events and their impact on the prices of financial assets.[45] Information extracted by the AI system from live news broadcasts is used in making investment decisions.
There are products that use AI to help people manage their personal finances. For example, Digit is an AI-based app that automatically helps consumers optimize their spending and savings based on their personal habits and goals. The app can analyze factors such as monthly income, current balance, and spending habits, then make its own decisions and transfer money to a separate savings account.[46] Wallet.AI, a growing San Francisco startup, creates agents that analyze data generated by consumers as they interact with smartphones and social media, in order to inform consumers about their spending.[47]
Automated robo-advisors are becoming increasingly widely used in the investment management industry. Automated systems provide financial advice and guidance in portfolio management with minimal human intervention. This class of financial advisors operates on algorithms designed to automatically grow a financial portfolio in line with clients' investment goals and risk tolerance. It can adjust to real-time market changes and calibrate the portfolio according to the client's wishes.[48]
The online lender Upstart analyzes vast amounts of consumer data and uses machine learning algorithms to build credit risk models that predict the probability of default. Its technology is being licensed to banks so they can use it to assess their own processes[49].
ZestFinance developed the Zest Automated Machine Learning (ZAML) platform specifically for credit underwriting. This platform uses machine learning to analyze tens of thousands of traditional and non-traditional variables (from purchase transactions to the way a client fills out a form) used in the credit industry to evaluate borrowers. The platform is especially useful for assigning credit scores to clients with limited credit history, such as millennials[50].
The use of AI allowed "Sberbank" to earn an additional $700 million in 2019, with the amount planned to reach $1 billion in 2020[51].
The use of AI is an important trend in the development of advanced battlefield and weapons control systems[52].
AI makes it possible to provide an optimal, threat-adaptive selection of sensor and weapon combinations, coordinate their joint operation, detect and identify threats, and assess enemy intentions[52]. AI plays a significant role in implementing tactical augmented reality systems. For example, AI enables the classification and semantic segmentation of images, and the localization and identification of mobile objects for effective target designation[52].
On March 1, 2021, the National Security Commission on Artificial Intelligence[en])[53] sent a report to the President and Congress recommending against a worldwide ban on the use of autonomous AI-based weapons systems (see also combat robot ). The report states that the use of AI will make it possible to "reduce decision-making time" in cases where a human is unable to act quickly enough. The Committee also expressed concern that China and Russia are unlikely to comply with a treaty banning the military use of AI[54].
China
According to the US Department of Defense, China has decided to develop methods for integrating AI into future weapons systems. The Academy of Military Sciences of China has been tasked with implementing this program by combining the efforts of the defense industry and private companies[55].
British intelligence services will combat Russian fake news using artificial intelligence that will recognize "troll factory" activity. According to the UK Government Communications Headquarters, artificial intelligence will fight fake news by cross-checking data against reliable sources, detecting image and video manipulation, and blocking suspicious bots[56].
Robots have become widespread across many industries and are often used for work considered dangerous for humans. Robots have proven effective in jobs involving repetitive routine tasks that can lead to errors or accidents due to declining concentration over time. Robots have also found wide application in work that people might find degrading.
In 2014, China, Japan, the United States, the Republic of Korea, and Germany together accounted for 70% of global robot sales. In the automotive industry, a sector with a particularly high degree of automation, Japan had the highest density of industrial robots in the world: 1,414 robots per 10,000 employees.
Artificial neural networks, such as the Concept Processing technology in EMR software, are used as clinical decision support systems for medical diagnosis.
Other tasks in medicine that could potentially be performed by artificial intelligence and are beginning to be developed include:
Currently, more than 90 startups based on the application of AI are operating in the healthcare industry
Another application of AI is in human resource management and recruiting. There are three ways AI is used for human resource management and hiring specialists. AI is used to screen resumes and rank candidates according to their level of qualification. AI is also used to predict a candidate's success in given roles through job-matching platforms. Finally, AI is used to create chatbots that can automate repetitive communication tasks.
Typically, the resume screening process involves analyzing and searching for information in a resume database. Startups such as Pomato are building machine learning algorithms to automate resume review processes. The Pomato AI system[62] is aimed at automating the review of technical applicants for positions at technical firms. Pomato's AI performs more than 200,000 calculations on each resume within seconds, then designs its own technical interview based on relevant skills.
From 2016 to 2017, the consumer goods company Unilever used artificial intelligence to screen all entry-level employees. Unilever's AI used neuroscience-based games, recorded interviews, and analysis of facial and speech cues to predict a candidate's success at the company. Unilever partnered with Pymetrics and HireVue to create a new AI-based analysis system and increase the number of candidates considered from 15,000 to 30,000 within a single year. Unilever also reduced application processing time from 4 months to 4 weeks and saved more than 50,000 hours of recruiters' time.
From resume screening to neuroscience, speech recognition, and facial analysis … it is clear that AI has an enormous impact on the field of human resource management. One of the advances in AI is the development of recruiting chatbots. TextRecruit released Ari (Automated Recruiting Interface). Ari — a suite of recruiting chatbots designed to conduct two-way text conversations with candidates. Ari automates posting job openings, advertisements, screening candidates, scheduling interviews, and developing candidates' relationships with the company as they move through the recruiting process. Ari is currently offered as part of TextRecruit's engagement platform.
Although the evolution of music has always been affected by technology, artificial intelligence has, through scientific advances, made it possible to imitate, to some extent, human-like composition.
Among the notable early efforts, David Cope created an AI called Emily Howell, which managed to become well known in the field of algorithmic computer music. The algorithm underlying Emily Howell is registered as a US patent.[63]
Other developments, such as AIVA (Artificial Intelligence Virtual Artist), focus on composing symphonies, mostly classical music for films. This development achieved fame by becoming the first virtual composer to be recognized by a professional music association.[64]
Artificial intelligence can even create music suitable for use in medical settings; Melomics uses computer-generated music to relieve stress and pain.[65]
Moreover, initiatives such as Google Magenta, run by the Google Brain team, want to find out whether artificial intelligence is capable of creating compelling art.
At the Sony CSL research laboratory, their Flow Machines software creates pop songs by studying musical styles from a huge database of songs. By analyzing unique combinations of styles and optimization methods, AI can compose music in any existing style.
In December 2020 in Russia, as part of the AI Journey conference (organized by Sberbank, moderated by Alexander Vedyakhin), Russian performers Zivert, Rakhim, Egor Ship, and Dania Milokhin performed together with artificial intelligence[66].
The company Narrative Science makes computer-generated news and reports commercially available, including summaries of sporting events based on statistical data from the game in English. It also produces financial reports and real estate analysis. Similarly, the company Automated Insights generates personalized recaps and previews for Yahoo Sports Fantasy Football. It is projected that by 2014 the company will produce a billion stories a year, compared to 350 million in 2013[67]. Leading media companies, such as Associated Press, Forbes, The New York Times, Los Angeles Times, and ProPublica, have begun automating news content. This has given rise to the concept of automated journalism[68].
Echobox is a software company that helps publishers increase traffic through the "smart" placement of articles on social media platforms such as Facebook and Twitter. By analyzing large volumes of data, the AI learns how specific audiences respond to different articles at different times of day. It then selects the best stories to publish and the best time to publish them. It uses both historical data and real-time data to understand what has worked well in the past, as well as what is currently trending on the Internet.
Another company, called Yseop, uses artificial intelligence to turn structured data into intelligent commentary and recommendations in natural language. Yseop can write financial reports, executive summaries, personalized sales or marketing documents, and much more at a speed of thousands of pages per second and in several languages, including English, Spanish, French, and German[69].
Boomtrain — another example of AI that aims to learn how best to engage each individual reader with precisely the articles — sent through the right channel at the right time — that will be most relevant to the reader. It is as if you hired a personal editor for each individual reader to pick out the best articles just for them.
There is also a possibility that in the future AI will write literary works. In 2016, a Japanese AI wrote a short story and nearly won a literary award[70].
Artificial intelligence is implemented in automated online assistants, which can be regarded as chatbots on web pages. This can help businesses reduce the costs of hiring and training staff. The core technology for such systems is natural language processing. Pypestream uses automated customer service for its mobile application, designed to simplify communication with customers[71].
Currently major companies are investing in AI to handle problem customers in the future. In its most recent version, Google analyzes human speech and converts it to text. The platform can identify angry customers through features of their speech and respond accordingly[72].
Many telecommunications companies use heuristic search in managing their staff; for example, BT Group deployed heuristic search in a scheduling application that provides work schedules for 20,000 engineers.
High hopes are placed on the use of artificial intelligence systems for managing 6G cellular networks[73].
In the 1990s, the first attempts were made to mass-produce home-oriented types of basic AI for education or leisure. This advanced significantly with the digital revolution and helped people, especially children, become familiar with various types of AI, in particular in the form of tamagotchis and virtual pets, the iPod Touch, the Internet, and the first widely distributed robot, Furby. A year later, an improved type of home robot was released in the form of Aibo, a robotic dog with intelligent features and autonomy.
Companies such as Mattel are creating a range of AI-enabled toys for children as young as three years old. Using proprietary AI systems and speech recognition tools, they can understand conversations, give intelligent responses, and learn quickly.[74]
AI is also used in the gaming industry, for example, video games use bots that are designed to play the role of opponents where humans are unavailable or undesirable. In 2018, researchers from Cornell University created a pair of generative adversarial networks and trained them on the shooter game Doom. During training, the neural networks identified the basic principles of level design for this game, after which they became able to generate new levels without human assistance[75].
Fuzzy logic controllers have been developed for automatic transmissions in cars. For example, in 2006 the Audi TT, VW Touareg, and VW Caravelle used the DSG transmission, which is based on fuzzy logic. A number of Škoda models (Škoda Fabia) currently also include a fuzzy-logic-based controller.
Today's cars now have AI-based assistance features, such as self-charging and advanced cruise control tools. AI is used to optimize traffic management applications, which in turn reduces waiting time, energy consumption, and harmful emissions by as much as 25 percent.[76] Fully autonomous vehicles will be developed in the future. AI in transport is expected to provide safe, efficient, and reliable transportation while minimizing harmful impact on the environment and society. The main challenge for the development of this AI is the fact that transport systems are inherently complex systems involving a very large number of components and different parties, each of which has different and often conflicting goals.[77]
In June 2019, a hardware-and-software system operating on computer vision technology was tested on the ChME3-1562 diesel locomotive assigned to Losta depot on the Northern Railway. In the event of danger (an incorrectly set switch, an obstacle on the track, a prohibitive traffic-light signal), the system first gives the driver a light-and-sound signal and then applies the brakes. [78]. The system, designated PAK-PML (hardware-and-software complex for locomotive driver assistance), uses artificial intelligence, accumulating data on trips already made and using it to assess the situation. In early September 2020, a trial run began at Losta station with two ChME3 locomotives now equipped with PAK-PML. The run is part of the pilot project of JSC "RZD" (Russian Railways), "Deployment of Technical Vision Technology for the Control and Monitoring of Rolling Stock." In turn, this project is an important stage of the larger corporate project "Digital Locomotive"[79].
Various AI tools are also widely used in security, speech and text recognition, data mining, and email spam filtering. Applications are also being developed for gesture recognition (machines understanding sign language), individual voice recognition, global voice recognition (from multiple people in a noisy room), and facial recognition for interpreting emotions and nonverbal cues. Other applications — robotic navigation, obstacle avoidance, and object recognition.
Combining artificial intelligence with experimental data has accelerated the creation of a new type of metallic glass 200-fold. The glassy nature of the new material makes it stronger, lighter, and more corrosion-resistant than modern steel. The team, led by scientists from the Department of Energy's SLAC National Accelerator Laboratory, the National Institute of Standards and Technology, and Northwestern University in the US, reported cutting the cost of discovering and improving metallic glass to a fraction of the previous time and expense. As team representative Apurva Mehta[80] said, "We were able to make and screen 20,000 candidates in a single year"[81].
In February 2021, the United States conducted tests of artificial intelligence in "two against one" air combat. The new phase of testing, called Scrimmage 1, was conducted at the Johns Hopkins University Applied Physics Laboratory. In this air battle, two AI-controlled F-16 Fighting Falcon fighters operated as a group and fought against one identical aircraft. During this new phase of testing, the neural network algorithms conducted not only close-range maneuvering air combat, but also operated at a distance from the adversary, detecting it with radar and striking it with missiles from afar[82].
Comments