Artificial Intelligence is already a major part of human life. There are different levels that AI can have. The one that we are most familiar with is the lower level AI that uses data and algorithm to schedule our news and social media feed. This type of AI can even play chess better than us. To other aspects of our lives are relying on this lower level of AI like organizational structures and facilities, and shaping the future of stock trading. The next step up is Artificial Intelligence for general purpose. This is an intelligence that are built to come close to human intelligence in the near future. What would happen to our world once we have successfully created an intelligence to compete and be on par with ourselves?
If you need assistance with writing your essay, our professional essay writing service is here to help!
The general purpose AI uses reinforcement learning and neural networks to imitate the learning process of the human brain. AI creators say that the algorithm can learn without alteration or instruction. Deepmind is a British company that created artificial intelligence for general purpose. It was founded in September 2010 and was acquired by google for 500 million dollars. Facebook also tried to buy Deepmind technology, but Google beat them to it. Elon Musk who often speaks publicly on the danger of AI to our world, is also an investor of the company, just to keep an eye on it. Deepmind started in 2010 by Demis Hassabis, who was a chess prodigy. He reached master level at age 13. Then went on to be a successful programmer at the University of London, where he met the co-founder of deepmind in 2010. The goal of Deepmind is to combine the best techniques of machine learning and neuroscience to build powerful general purpose learning algorithms. In other word, they wanna formalize intelligence, find out what it is and how it works, and also to understand the human brain, as our mind is a mystery.
In an experience, the developer let the machine play the game Space Invader. First, they showed the machine the data for the first time, the machine learn how to play the game on its own. They leave the machine learn on its own overnight for about 8h. When they come back in the morning, the machine was at superhuman level. The machine now can play Space invader better than any human can. After understanding how to play a few games, it is able to transfer this knowledge and understanding across more games, and then move on to other more advanced and complicated application. General purpose AI algorithm has unlimited application. Things like improve the speed of advancement in healthcare and science to help human expert achieve more. Online customer service, finance, computer vision, general computer science, news and publishing and writing could also be huge. General purpose digital neural that use reinforcement learning can be the groundwork for imaginable thing in the near future.
For a long period of time of our existence, we have advanced so much with technology, improved life quality. But AI is different. AI is the first thing human have created that tend to function in a way that we cannot predict. The fear of the unknown is mostly drive the AI anxiety. For a long time, people weren’t taking AI seriously, because it seems like the progress is really slow, but it had proven its unpredictability and overtook human expectation. One of the methods that AI can use to achieve the outcome is by giving itself an ability for self improvement, thus following the law of accelerating. Stephen Hawking, Elon Musk have shown their worried about it. But Paul Ellen, co-founder of Microsoft had a different view. He discussed in an interview that the possibility that AI can be human aid fixing fundamental problems in our planet that and allowing us to have technology and infrastructure that could otherwise take centuries if we try to figure it out ourselves. (AI2)
Could AI be able to reach human’s level of intelligent and could be smarter than us? A survey was carried out asking some of the world leading experts on AI by which year do you think there is a 50% probability we will achieve human level machine intelligence? We defined human level here as the ability to perform almost any job as flawless as an adult human level and the median answer was around the year of 2030 to 2060 depending on which group of experts. (Ryan) Now, it can happen much later or much sooner, nobody really know, but what we do know is the ultimate limit of machine is far outside the limit of biological issue. This comes down to physics, a biological neuron fire 200 HZ, 200 times/s, 100m/s top (aiimpacts.org). A computer, signal can travel at the speed of light. There are also size limitations like the human mind is contain within cranium, while a computer can be a size of warehouse or larger.
Andra Keay, the managing director for Silicon Valley robotics predicted that in the time of 5-10 years time, robotic is going to disappear from our awareness because we will be so familiarize with it, it is going to be what all of our smart product has (Deskos). But first, we need to understand how far AI and robotic has become. Warehouse and factories are being run by robots. Cars and planes are operating by themselves. Robots are doing fine detailed work and response to every move. They are entering the service factor, replacing fast food workers and baristas. Robots are replacing jobs like security guards with self-driving surveillance vehicles. It’s impossible to see such futuristic technology without wondering about the impact on jobs. And endless array of robots are entering our everyday lives. Pair that with advances in AI, machine learning, 3D printing, and the changes are hard to comprehend.
As of the present day, given all the knowledge of the advancement of AI, the biggest job killers for people are driverless vehicle which are already driving on our road. World wide, 800 million jobs could be automated by 2030. In the US, up to 73 Mil jobs are at risk. Changes like these have taken place in history. Like the industrialization of agriculture, only this time, it is on a much bigger and faster scale. In 2017, Uber bought 24,000 self driving cars from Volvo and put them to work (Pollard). It’s not hard to picture Uber and Lyft based jobs collapsing as self-driving cars arrive. Everyday, Lyft and Uber drivers are giving the data needed to automate. When we book a ride on an Uber or Lyft with our smartphones, we are unwillingly teaching AI how to do their jobs. Three percent of the American workforce drives for a living, and that does not includes hotel along highways, truckstops, restaurants and other support jobs. (whitehouse.gov). We cannot emphasize enough on the importance of entry level, middle class job for millions of Americans. Trucking is the most common jobs in 29 US states. There are 1.3 to 1.7 million truck drivers at risk for automation, add in taxis, chaffers and delivery trucks and that number rises to 3.1 millions (whitehouse.gov). The world health organization estimated that 1.25 mil people die each year in traffic accident. But self-driving vehicle can reduce the number of accidents by 90%. Which translate to 29,447 a year, 300,000 lives per decade. (LaFrance)
Even safe and stable white collar jobs like doctor, bankers, lawyers are at risks. We already use services like Turbotax to replace tax advisors because it is within reach and significantly convenience, and cheap. Amazon Alexa and Apple Siri has taken secretarial tasks. AI is making ways on the legal profession, finance sector and the medical world. Developers are working on therapy and counselling with AI. “We need to start thinking about what’s gonna happen when large numbers of human white collar workers are replaced” Neil Jacobstein Chair of AI and Robotics at Singularity University said. Movo is a research robot launching in 2018 and other robots have already made advances in medicine. Some surgeries outcome were improved by using assisted AI. The variation in procedure due to humans’ mistakes were reduced and hopefully eliminated. In fall of 2017, Maastricht University Medical Center in the Netherlands used an AI assisted robot to suture blood vessels that are as small as .0012 inch. (Britt). Baxter Robot is an automation robots that do jobs that are boring and repetitive, and unsafe. The service industry is also transforming. From line cook to baristas, to servers. Pizza Hut announced driverless delivery and claimed that “it could actually create more jobs by opening the pool of drivers to those who do not own vehicles, or do not have a driving license. They act more as servers, focusing on hospitality” (Higgins). Johannes Moenius, an economic professor at the University of Redlands said that approximately 58% of jobs in the Los Angeles area and other service industry heavy cities like Las Vegas area are at risk of automation by the year 2035. These cities are more prone to automation because of the high number of service job, and few robots are already working in place of humans. Casino employment peaked in 2001 peaked at 60,000 and by 2016, it dropped to less than 41,000. In a factory in 1980s, it took 25 people to produce $1 million output. Today, just five people are needed for the same result. (Mason).
Replacing human laborers with automatons is not entirely unethical. AI can replace the dangerous and menial jobs, they can also the complex, fast pace, extremely precise, and knowledge heavy jobs. The ethical issue in this situation is complex, because it doesn’t just involve the relationship between humans and robots, but also between humans and the future of us. Is it ethical to stop improvement? Would it be ethical for us to try and stop Baxters, Google cars, warehouse robots, and assisted medical AI? Up until this point in time, people were all we used to have, so all the jobs have been done by people. Human can continue to exceeds our being by optimize the use of robotics and machine intelligence. Machine intelligence is the last invention that human will have to make. The machine will then be better at inventing than humans are, and they’ll be doing so in digital time scale. Cures for aging, space colonization, free and clean energy. All of it, an AI with superlogical maturity would be extremely powerful and effective to make science fiction into a reality.
On the other hand, intelligence does not come with value and a set of judgment like humans. It’s more relieving to assume that anything intelligent will just naturally develop human value like empathy or respect for life. This can’t be guaranteed because ethical value are based purely on human principle. Given that we as humans find it impossible to agree among ourselves of what is right and wrong. For example areas like border control, euthanasia, abortion and more. How can human possibly program a computer with a knowledge that we ourselves can’t even agree among each other.
A machine intelligence might know right from wrong. But can it learn abstract model of how the world works, and what fairness is? For an AI, when something is not fair, it interprets that thing as not efficient, and not the most cost effective. We needs human to make these important decision. When we set certain goals, the AI learning algorithm can shortcut those goals to reach the goal in the quickest and most cost effective manner, but not necessary the most fair. For example, in the medical field, if the goal of the system is to treat as many people as it can. The AI would give more priority to people with simpler injuries, like treating a small burn, or repairing a small joint sprained, so the number of people helped could go up higher. The case that needed a 10-hour operation is put to the side because there were 10 other cases who could take only 45 minutes to operate quicker and cheaper. So therefore, it chose 10 people get treated rather than one. It is purely a number game for the AI. Treating people is incredibly complicated because not only we are dealing with risks, we are also dealing with many factors including age, the severity of the illness, lifestyle, income, diet of the patients. All these things come into the factors that play into how to treat people. AI could make a wrong decision that is unethical, but according to the data given, it is the right answer. In addition, machine learning doesn’t understand biased rule. For example, they have all the data and characteristics of a case to form a treatment, but there is something that is missing, that can actually change the characteristics of the whole dataset like the value of life, the value of another family member that requires human intuition.
In conclusion, if we try to stop to machine intelligence development, we are stopping human progress for fear of the unknown. It is almost like refusing to eat for fear of food poisoning. It is counter productive for human progression. Humans have made it to this point by determination and spare no efforts until everything works. On the other hand, science without restrictions leads to major repercussion. Research into the field of AI, like most everything else in life, should be done with serious consideration of ethical values. Policy makers and government needs strictly regulate AI in order to protect humans and progress along with the advancement of AI. In the world of AI, machines would still be our servants, and they would help us solve all of the matters that been troubling us.
- “About Us.” DeepMind, deepmind.com/about/.
- AI2. Paul Allen Discusses AI2 and the Future of AI. YouTube, YouTube, 4 June 2014, www.youtube.com/watch?v=iq3Fgh6Ovi0.
- Artificial Intelligence, Automation, and the Economy.” Whitehouse.gov/, National Science and Technology Council, www.whitehouse.gov/sites/whitehouse.gov/files/images/EMBARGOED%20AI%20Economy%20Report.pdf.
- Britt, Phil. “How AI-Assisted Surgery Is Improving Patient Outcomes.” Robotics Business Review, Robotics Business Review, 15 June 2018, www.roboticsbusinessreview.com/health-medical/ai-assisted-surgery-improves-patient-outcomes/.
- Deskos, Nicolas. “Artificial Intelligence: Should We Feel Fear or Hope?” Europe’s Leading Branding Conference – OnBrand ’19, onbrand.me/magazine/artificial-intelligence-fear-or-hope/.
- Higgins, Tim. “Pizza Delivery Gears Up for a Driverless Era.” The Wall Street Journal, Dow Jones & Company, 26 June 2018, www.wsj.com/articles/pizza-delivery-may-be-entering-a-new-era-1530029087.
- LaFrance, Adrienne. “Driverless Cars Could Save Tens of Millions of Lives This Century.” The Atlantic, Atlantic Media Company, 23 June 2016, www.theatlantic.com/technology/archive/2015/09/self-driving-cars-could-save-300000-lives-per-decade-in-america/407956/.
- Mason, Melanie. “Millions of Californians’ Jobs Could Be Affected by Automation – a Scenario the next Governor Has to Address.” Los Angeles Times, Los Angeles Times, www.latimes.com/projects/la-pol-ca-next-california-work/.
- “Neuron Firing Rates in Humans.” AI Impacts, AI Impacts, 17 Apr. 2015, aiimpacts.org/rate-of-neuron-firing/).
- “Neil Jacobstein on the Future of AI.” 1761eac3b1967fa6e4cafb6af110abcc-1, NextBigFuture.com, 15 Feb. 2018, www.nextbigfuture.com/2017/08/neil-jacobstein-on-the-future-of-ai.html.
- Pollard, Niklas. “Volvo Cars to Supply Uber with up to 24,000 Self-Driving Cars.” Reuters, Thomson Reuters, 20 Nov. 2017, www.reuters.com/article/us-volvocars-uber/volvo-cars-to-supply-uber-with-up-to-24000-self-driving-cars-idUSKBN1DK1NH.
- Ryan, Kevin J. “Elon Musk (and 350 Experts) Predict Exactly When Artificial Intelligence Will Overtake Human Intelligence.” Inc.com, Inc., 6 June 2017, www.inc.com/kevin-j-ryan/elon-musk-and-350-experts-revealed-when-ai-will-overtake-humans.html.