Current Affairs

ARTIFICIAL INTELLIGENCE—IT CAN DELIVER; IT CAN DESTROY TOO

on
August 12, 2020
1. PREAMBLE
1.1. It has been a dream for a long time for humans to create machines that have the ability to take decisions on their own, much faster than the human brain can ever do.
1.2. We have progressed a lot in this field and right now we are in a situation where we are constantly surrounded by examples of “narrow AI” or “weak AI”. These are machines or programs that can carry out a simple or complex specific task by itself and that too very efficiently.
1.3. For example, a self-driven car can drive you from home to work, taking in account the traffic conditions, environmental factors etc. There are a lot of digital assistants in the market like Amazon Alexa or Apple Siri that make our lives easier.
1.4. But there is always a constant debate going on since the inception of the idea of artificial intelligence. There are divided opinions on whether,–
1.4.1. humans are digging their own graves by creating intelligent machines or
1.4.2. they are contributing in catapulting human civilization to greater heights with it.
2.BENEFITS OF AI
2.1. Increase in work efficiency
AI-powered machines are great at doing a particular repetitive task with amazing efficiency. The simple reason is that they remove human errors from their tasks to achieve accurate results every time they do that specific task.
2.2.Moreover, such machines can work 24X7, unlike humans. Thus, they eliminate the need to deploy two sets of employees working in day and night shifts to work on important tasks. For example, AI-powered chat assistants can answer customer queries and provide support to visitors every minute of the day and boost the sales of a company.
2.3.Work with high accuracy
2.4.1.Scientists are working to teach artificial intelligence powered machines to solve complex equations and perform critical tasks on their own so that the results obtained have higher accuracy as compared to their human counterparts.
2.4.2.Their high accuracy has made these machines indispensable to work in the medical field particularly, owing to the criticality of the tasks. Robots are getting better at diagnosing serious conditions in the human body and performing delicate surgeries to minimize the risk of human lives.
2.5. Reduce cost of training and operation
2.5.1.AI uses machine learning algorithms like Deep Learning and neural networks to learn new things like humans do. This way they eliminate the need to write new code every time we need them to learn new things.
2.5.2.Since they learn MUCH FASTER than humans, the cost of training robots would become much lesser than that of humans.
2.5.3.Also, machines don’t take breaks and can perform the same mundane task again and again without any downtime or change in results.
2.6. They Improve process performance
2.6.1.The best part about AI-powered machines being deployed for work is that they let us gather humongous amounts of data related to their work. Such data can be processed to gather deep insights into the processes with quantitative analysis so that we can optimize them even further.
2.6.2.Machine learning abilities of AI machines are increasing further and further to do even the analysis by themselves.
2.7. THE GPT 3 IS HERE!
2.7.1.Made by Open AI Software, it is generally agreed that the GPT-3 is probably the best non-human writer ever!
2.7.2.Given a sentence and asked to write another like it, the software can do the task flawlessly.
2.7.3.GPT-3 stands out because it has been trained on more information – about 45TB worth – than anything else.
2.7.4.Because the software can remember each and every combination of words it has read, it can work out – through lightning-fast trial-and-error attempts of its 175bn settings – where thoughts are likely to go.
2.7.5. Remarkably it can transfer its skills: trained as a language translator, GPT-3 worked out it could convert English to JavaScript as easily as it does English to French.
2.7.6.THE FLIP SIDE!
(i)GPT-3 doesn’t know what it is doing; it is unable to say how or why it has decided to complete sentences; it has no grasp of human experience; and cannot tell if it is making sense or nonsense.
(ii)What GPT-3 represents is a triumph of one scientific paradigm over another.
2.8.HOW AI MACHINES STARTED BEATING CHESS GRANDMASTERS
2.8.1.Once machines were taught to think like humans. They struggled to beat chess grandmasters.
2.8.2.Then they began to be trained with data to, as , “discover like we can” rather than “contain what we have discovered”.
2.8.3.Grandmasters started getting beaten. These days they cannot win.
3. THE RISKS
3.1. AI can emulate the racist and sexist biases of the data it swallows. In an era of fake news, GPT-style devices could become weapons of mass destruction: engaging and swamping political opponents with divisive disinformation.
3.2. AI can encompass anything from Google’s search algorithms to IBM’s Watson to autonomous weapons.
3.3. Autonomous weapons are artificial intelligence systems that are programmed to kill. In the hands of the wrong person, these weapons could easily cause mass casualties.
3.3.1. Moreover, an AI arms race could inadvertently lead to an AI war that also results in mass casualties.
3.3.2. To avoid being thwarted by the enemy, these weapons would be designed to be extremely difficult to simply “turn off,” so humans could plausibly lose control of such a situation.
3.4. The AI is programmed to do something beneficial, but it develops a destructive method for achieving its goal:
3.4.1. This can happen whenever we fail to fully align the AI’s goals with ours, which is strikingly difficult.
3.4.2. If you ask an obedient intelligent car to take you to the airport as fast as possible, it might get you there chased by helicopters and covered in vomit, doing not what you wanted but literally what you asked for.
3.4.3. If a super intelligent system is tasked with an ambitious geoengineering project, it might wreak havoc with our ecosystem as a side effect.
3.4.4. Worse, it will view human attempts to stop it as a threat to be met, just like our white blood cells attacking other cells thinking they are the enemies of the body, in autoimmune diseases like arthritis!
4. THE PROBLEM IS NOT ONE OF MALEVOLENCE BUT COMPETENCE!
4.1. The concern about advanced AI isn’t malevolence but competence.
4.2. A super-intelligent AI will be extremely good at accomplishing its goals, and if those goals aren’t aligned with ours, we have a problem.
4.3. You’re probably not an evil ant-hater who steps on ants out of malice, but if you’re in charge of a hydroelectric green energy project and there’s an anthill in the region to be flooded, too bad for the ants.
4.4. A key goal of AI safety research is to never place humanity in the position of those ants!
Ramdas Iyer
12.08.2020
Wednesday
TAGS
RELATED POSTS

LEAVE A COMMENT