Almost 20% of experts in the Artificial Intelligence industry believe that AI will someday pose a threat to humanity, a study done by Oxford University has revealed.
Of the 550 respondents, 24% believed that AI systems will have ‘Extremely good’ effects on humanity, 28% thinks it will bring ‘On balance good,’ 17% are ‘More or less neutral,’ 13% feels it will cause ‘On balance bad,’ and 18% believes it will bring ‘Extremely bad’ consequences.
Experts predict that AI will reach an overall human ability level between 2040 and 2075, and that’s when the concept of singularity will come into reality. The concept of singularity, which is the basis of the fears around AI, is that point in the future when it will be impossible to predict the growth of AI because it has become a super-intelligent being, with roughly the same human level of ability. We’re talking about AIs being able to think and decide for themselves.
And this is not the first time Oxford has given this warning against AI. In February 2015, Oxford University also released a list of local threats that pose risks to human civilization and the report warned that there is a possibility that AI might cause mankind’s extinction one day, deliberately.
These studies from Oxford University are not the first to tackle the dangers of AI. In an interview with the BBC in 2014, famous physicist Stephen Hawking was quoted saying:
“The development of full artificial intelligence could spell the end of the human race. It would take off on its own, and re-design itself at an ever-increasing rate. Humans, who are limited by slow biological evolution, couldn’t compete and would be superseded.”
Tesla and SpaceX innovator Elon Musk also warned against the risks of AI development. In an interview with CNBC, Musk revealed that the reason he was investing heavily in AI companies, such as DeepMind and Vicarious, is so that he can keep an eye on what is going on in the AI industry. Microsoft leader, Bill Gates, also expressed his concern about AI and superintelligence. However, Gates also believes that AI can be used as better software to improve human lives.
Also Read: Virtual Reality or Artificial Intelligence
The debate regarding the effects of artificial intelligence on human civilization has been going on for quite a long time. AI and tech experts are divided about the issue, with some believing that AI will bring extinction upon the human race while some think that AI can be leveraged to improve industries, such as agriculture, energy resources, among others.
What Is Artificial Intelligence?
Artificial intelligence, commonly known as AI, refers to the branch of computer science that deals with creating smart machines with the ability to perform tasks that typically require human intelligence. It is an interdisciplinary study with multiple approaches, but recent progress in AI is creating huge waves in various fields.
Decision-making, speech recognition, translation, and visual perception are some of the factors that determine human intelligence, that can also be possessed by artificial intelligence. Another feature that makes AI powerful is the ability to learn from experience, also known as machine learning.
Artificial intelligence is a combination of concepts and technologies and has a lot of applications today. AI is often used for video games, software simulation, robotics, self-driving cars, smart homes, and as personal assistants (Cortana, Siri, and Alexa). Even Google’s search algorithm is powered by AI. You can see its applications in all aspects of our lives.
AI is composed of two subsets of techniques: machine learning and deep learning. Machine learning uses statistical concepts to enable smart machines to improve with experiences. Deep learning, on the other hand, makes the computation of multi-layer neural networks possible.
Artificial intelligence is a rapidly growing industry, with global spending projected to increase by 50% compounded annually and reaching $57.6 billion by 2021. In many advanced economies and emerging economies, artificial intelligence has become a staple growth factor in the workplace and businesses. It is also widely used in homes, medical facilities, government offices, and even our everyday lives.
There are many AI applications that help make our everyday lives more efficient and less stressful. What tech experts actually fear are those applications that play a crucial role in ensuring our safety and our resources. For example, if an artificial intelligence system that manages the country’s intel gets hacked or goes wonky for some reason, the consequences of that incident can be disastrous. Not only will it affect the country involved, but other related countries as well.
What Makes Artificial Intelligence Dangerous?
The era of Terminator-like super machines are still 20 to 50 years away into the future, but the social, political, financial, and regulatory components of artificial intelligence are so complicated and far-reaching that we need to scrutinize them now so we know how to handle them when they become a reality.
Aside from preparing for an AI-powered future, it is also critical to look at how AI can pose risks in its current form. This article will discuss the top five AI applications that are capable of bringing more harm than good.
1. Smart Weapons
Incorporating artificial intelligence in weaponry is one of the most dangerous applications of this technology. We have heard of machine-operated nuclear attacks or warfare programmed by AI in the news, but the danger is more tangible than we can possibly imagine. Although AI can improve the weapon’s accuracy and efficiency, it can also pose risks, not only to the targets but to the people overseeing the weapons as well.
With the various countries competing in the use of AI in the nuclear arms race, it is unimaginable what another World War would be like, in case one breaks out. According to President Vladimir Putin of Russia, artificial intelligence comes with enormous opportunities, but the threats are also difficult to predict. He even stated that whoever becomes the leader in AI will most likely become the ruler of the world.
Imagine fighting enemies with tanks and ships controlled remotely by AI or drones surveying the topography with no pilot. With no human lives to consider, autonomous weapons can operate unhindered.
However, aside from fearing that autonomous weapons might gain a mind of their own, it would be more practical to think of how these weapons can be used by people who don’t value human life. For example, a terrorist can easily plant a bomb inside a robot, use its built-in machine visual feature to track down the target, then detonate on site. The attack will be hard to track down because of the absence of human component, and it will also make the attackers bolder and creative in deploying their acts of violence.
2. AI-Powered Online Attacks
One of the benefits of using AI is automation. You can use it to reproduce tasks that previously required human labor. If used well, AI can increase efficiency and productivity a hundredfold. But if it falls in the hands of the wrong people, it means more damage to the public. Some online attacks, for instance, can be automated and cause more damage to more people.
Let’s take spear phishing as an example. In this kind of attack, the hacker sends messages designed to trick users into giving up their security credentials, financial information, or other sensitive data. With an AI-powered software, hackers can now send more spear-phishing emails to more contacts, with no extra cost. Phishing emails, by themselves, are dangerous enough. Imagine what an automated AI phisher can do.
The same concept applies to other online attacks, such as ransomware, crypto-jacking, adware, and other types of online attacks.
3. Social Manipulation
One prime example of AI being used in manipulating people is Google. Its autonomous-powered algorithms provide Google with all the information about all users. Because of this, Google knows where we are, what we do, what we eat, what we like, and probably what we’re thinking.
If people only knew how much data Google collects, you’d probably think twice the next time you use Google to search for something. Another good example is social media. It may not be as powerful as Google, but the results are the same. So if you want to prevent Google and social media sites from accessing your data, deleting your cache using a PC cleaning software will ensure that there will be no data to collect from you.
Data gathered from Google and social media can be used by different entities in different ways. For example, marketers can use consumer’s data for target marketing. Politicians and organizations, on the other hand, can use this data to sway public opinion. Just look at the Cambridge Analytica scandal.
Investigations are still being conducted regarding Cambridge Analytica’s role in trying to sway the outcome of the 2016 U.S. presidential election, as well as the U.K.’s Brexit referendum, but issues like this illustrate the capability of AI for social manipulation.
By using AI-powered algorithms and matching personal data, AI can target the right people and spread propaganda in whatever format that is most convincing, regardless of whether the message is a fact or fiction.
4. Invasion of Online Privacy
Surveillance is a powerful data-gathering tool. You can now track and analyze every person’s activities, both online and in person. Cameras are everywhere and facial recognition algorithms allow AI to know who you are. This is what happened in China last December 2018 where an entire border region composed of largely Muslim Uighur minority, underwent a total surveillance state that can only be considered as oppressive.
China is using its wide network of CCTV cameras to implement its social credit system. This system will assign a personal score for every citizen based on how they behave. The system will be able to see what people are doing, both online and in your everyday life. All their activities are monitored and the scoring system factors these in when giving the score.
Although some users are tricking the system by installing Mac cleaning software or cleaner apps for Android to delete their online activities, they still can’t get away from real-life surveillance. This is clearly a violation of personal and online privacy, and AI is giving China the power to accomplish it.
5. Proliferation of Fakes
We have seen a lot of “deepfake” videos last year, and this issue reveals how easy it is to create fake video clips using AI. The “deepfake” videos included pasting of faces of celebrities onto porn clips without their consent. This also applies to weird movie mashups and possible political smear campaigns. Generative adversarial networks or GANs utilize two neural networks to create fake digital images known as “deepfakes.”
Although GAN’s ability to synthesize data patterns can be quite useful in animation, photo editing, and medicine, it can also be used in a lot of dangerous ways. Aside from overlaying celebrity faces on porn clips, you can also make politicians say whatever you want or forge someone else’s biometric data for malicious purposes. Because “deepfakes” are so realistic, it is hard to detect them manually, so we’ll probably have to use another AI for the job.
Artificial intelligence is a double-edged sword. It has endless applications, but so are the risks. Being aware of the dangers of AI is just the first step in preventing these potential risks from becoming real threats. It will take a concerted effort between governments, the scientific community, the AI experts, and the businessmen financing these technologies, to ensure that AI remains an asset, not a burden.