Professor of Oxford University Nick Bostrom believes that computer super mind poses a threat to our civilization. Other scientists, for example, Stephen Hawking, have also repeatedly expressed similar concerns about artificial intelligence. However, artificial intelligence technologies keep being advanced and updated. It means that the advantages of computer mind are worth risking human security.
Beneficent machines or why do we need artificial intelligence
Artificial intelligence implies a possibility of machines to think and learn as a human. The research field of AI is relatively young, as the first developments in this area started just in 1956; nevertheless, it does not stop people from implementing AI technologies in many activities today. This includes smartphones and gadgets, diagnostics of diseases, electronic document management, car industry, and housekeeping.
Development of artificial intelligence is intended to make the life of people more comfortable and convenient, while making performance more efficient. Many global scientists, futurologists, and thought leaders hold to this opinion. For example, CEO of FacebookMark Zuckerberg believes that “artificial intelligence will be changing our life only for better in the next five or ten years”.
AI technologies are steadily implemented in all possible areas of human activity, but several key areas can be distinguished, where computer mind is most effective.
Diagnostics. Artificial intelligence helps to improve healthcare and patients’ service rate. Today physicians can use IBMWatson Healthproduct and GoogleDeepMind Health, an intelligent assistant that helps to make a diagnosis. Besides, there are such systems as Zebra Medical Visionintended for visual examination of bowel and magnetic resonance tomography, and apps like Ada.
Using convolutional neural networks (CNN), Australian scientists have developed a program for melanoma recognition. Trials have shown that the system diagnoses melanoma more accurately than dermatologists do. Computer mind detected melanoma in 95% of cases, whereas experienced physicians just in 88.9%. In such a way, artificial intelligence can significantly improve cancer detection, which influences human life.
CCS Insightanalytics firm believes that AI will boost the quality level of medical services and make patients treatment more productive in the next five years.
Automation of routine processes. Artificial intelligence is designed to assist people in housekeeping and at work. Smart home systems and personal assistants such as Siri and Alexa simplify the daily routine: smart home systems ensure safety and automation of your swelling, while smart assistants draft timetable, browse the web, and systematize important information.
Using AI technologies, Amazon foresees customer needs, and Netflix selects a list of films that may interest users. Machine learning technologies lay the base for automation of business processes. For instance, Now Platform prevents down times for customers, assesses efficiency of IT, streamlines service backup, and conducts predictive analytics.
Optimization of production. Results of the research from Cifra show that the Russian industrial market of artificial intelligence will amount to $380million by 2021. All major enterprises use AI technologies today.
Computer mind optimizes operation of plants, conducts quality assurance, and reduces production costs. Machine learning methods are applied in aircraft engineering, car building, instrument engineering. Some enterprises use them partially (main decisions are taken and assessed by a human), and LG, for example, is planning to open a factory in 2023, where artificial intelligence will control all processes.
Security. Smart homes and smart cities will help people to not only reduce consumption of resources and optimize routine processes. Artificial intelligence systems are able to adapt to human habits, recognize faces and voices, use HD cameras, audio sensors to ensure security.
AI can evaluate the situation on its own and call for emergency services (in case of fire, housebreaking, and other accidents). Besides, smart technologies help to identify people during financial operations in particular.
Smart cities also provide high security level: face recognition systems track thieves, record malefactions. For instance, Hikvision, a company that produces video surveillance devices, uses AI technologies to increase the security level in towns.
Why artificial super mind is dangerous for humanity
Use of artificial intelligence brings obvious benefits to people, but at the same time raises a number of important questions and provokes reasonable concerns. What if the technologies will be used for cheating or fraudulent activities? What dangers may the creation of a super intelligent computer bring?
Philosopher Nick Bostrom, physician Stephen Hawking, magnate Bill Gates, and engineer Elon Musk support the idea of artificial intelligence posing a threat to humanity. Main areas that raise concerns include labor management, human safety, and personal data protection.
Reduction in the number of workplaces. Robots and AItechnologieshelp tostreamline operations at harmful and dangerous production facilities, where people are at risk of traumatizing or losing life. Innovative solutions help to replace people where boring, monotonous work is required, related to calculations and analytics. However, this also leads to the reduction in the number of workplaces and vanishing of some professions.
Research conducted by McKinsey Global Institute shows that 5% of specialties can be completely automated today, and in the future, robots will be able to fulfill functions of 60% of professions. Scientists have calculated that 57activity types will have disappeared by 2030. They include copywriters, analysts, librarians, financiers, delivery people, and waiters.
Nevertheless, experts forecast that new specialties will emerge. For example, the US consultancy firm Cognizant Technology Solutions thinks that 21 new specialties will appear in the next 10 years, including data broker, IT intermediary, and personal memory curator.
Incorrect use of technologies. Currently artificial mind is capable of learning and taking decisions, but the direction of work and end goals are still set by a human. Correspondingly, nothing stops standalone people and countries from making artificial intelligence a powerful weapon and instrument of control. In China, the government is actively using AI and IoT technologies to track citizens turning the country into a technological anti-utopia.
Besides, individuals can deploy artificial intelligence to harm other people. They are digital scammers and hackers. Specialist of the Centre for the Study of Existential Risk of Cambridge University Shahar Avin distinguishes several areas, where AI can be deployed with a negative purpose:
- such software as AlphaGo, which outcompeted a human in the board game Go, can be used by hackers to find code vulnerabilities;
- specially programmed smart bots can produce fake news and video to manipulate people’s consciousness;
- fraudsters can synthesize speech to bypass voice identification systems.
These and other similar concerns relate not to the threat that artificial intelligence poses as such, but the issue of control over the use of new technologies.
Rebellion of machines. Computer outperformed the best Go player back in 2016; IBM Watsonwon the intellectual quiz Jeopardy; artificial intelligence diagnoses melanoma more accurately than a human does. Does this mean that computer brain is already smarter than we are? What we should fear? Philosopher Nick Bostrom believes that super smart artificial intelligence might consider people an obstacle for achieving a goal (for example, to produce as many clips as possible) and eradicate humanity as a minor hurdle.
For now, such assumptions look like science fiction, as artificial intelligence is underdeveloped. Not a single intelligence systemoutperforms human brain, and machines lack human motivation and emotions.
To protect the world against potential threats of artificial intelligence, ethical principles of AI development are drafted as well as rules of secure use of new technologies.
23 principles of secure AI development
A list of Asilomar AI Principles was presented at Beneficial I conference back in 2017. Physicists, robot technicians, IT specialists, developers, and scientists participated in the creation of rules. Elon Musk and Stephen Hawking are among the people that signed the paper.
The document stipulates ethical principles of AI use including:
- aims of artificial intelligence systems should not contradict universal human values;
- the developer bears the responsibility for harm brought by artificial intelligence;
- AI systems capable of reproducing themselves should be strictly controlled.
In 2018, a group of human rights defenders from the Electronic Privacy Information Center (EPIC) developed 12 guidelines on robotics. These rules are also intended to regulate the use of artificial intelligence. After the publication was made, activists suggested that the US authorities officially adopted the rules: 200 experts and 50 organizations including the Federation of American Scientists signed the initiative.
Despite concerns and warnings, AI technologies keep being developed at an unbelievable pace. According to AI Indexannual report prepared by MIT, Stanford University, and Harvard University, the number of functional AI startups grew by 2.1times during the period from 2015to 2018in the US, and their total number grew by 1.3times. Venture investments in such enterprises have grown by 4.5(since 2013(general investments – by 2.1times).
Futurist Raymond Kurzweil thinks that computers will match human brain in terms of computational capability and even outperform it in 2040s. The expert feels optimistic and believes that humanity will start using smart nanorobots to help people combat diseases and aging in the upcoming decades.