Wouldn't it be nice to have Honda's Asimo wake us up for work, put the eggs in the frying pan and bread in the toaster, and send us off like something out of a science fiction story by Ray Bradbury? The friendly household robot that helps with the chores is one image of artificial intelligence (AI) that many of us can get behind.
Then there's the super-efficient AI in factories, on our computers at work, making repetitive work easier, minus the errors we humans inevitably make. It's easy to picture the robotic concierge, checking in guests at the front desk of a hotel or government building, and making sure no unauthorized people are around. Or perhaps the person who takes your order at the drive-in will no longer be a person, but a tireless machine, which will be faster, won't need a break and won't ever ask for a payrise.
AI is set to make our lives easier - sooner than you think, according to experts. But can AI take away our jobs? The big question is, can AI replace humans at everything we do? And what can be the real consequences for the future of humanity if AI does things better than us?
Experts say AI can replace humans in 45 years
Researchers from Oxford University's Future of Humanity Institute spoke to 352 experts in machine learning to see what they had to say about our future with AI. The unnerving forecast is this:
In ten years, AI will be better than us at translating languages, driving trucks and writing high school essays. If AI keeps growing the way it has so far, then machines will be taking our retail jobs by 2031 and be putting best-selling writers out of business by 2049. By 2053, surgeons will have to find other means of livelihood, because AI will outperform them. Some experts say that in the next 120 years - others say sooner - AI could take over all human jobs.
So, what does this mean for society? Will we all get to enjoy the benefits of having everything done for us by capable machines, while we take up our favorite leisure activities in the middle of the day? Not exactly.
What AI means for jobs
According to experts, AI will probably have the biggest impact on people who have lower income earning jobs. The Council of Economic Advisors (CEA), who advise the White House, have already considered robots in the workforce. People in low paying jobs are the most likely to be replaced with robots. According to the CEA, the chances of this happening is a whopping eighty-three percent. This forecast is scary because, according to 2014 figures from data compiled by Goldman Sachs, the majority of American jobs pay less than $20 an hour.
If the experts are right, millions of salespersons, cashiers, service workers in the food industry, office clerks, waitresses, waiters, laborers, cleaners, janitors and other blue collar workers are going to lose their jobs to machines. The report by CEA also shows that people who earn more are less likely to be replaced by AI. So, if you make more than $20 an hour, there's a thirty-one percent chance that a robot will be writing your articles in the future.
If you earn $40 an hour, however, there's only a four percent risk that a robot could replace you. Plus, there will always be a need for skilled freelancers. (You can safely apply to jobs on Freelancer.com without having to worry about robots taking your job in the near future.)
Given such a possible scenario may occur soon, President Obama's education policies make a lot of sense. Rather than encouraging a Universal Basic Income, to be paid to everyone to make up for large scale automation that people are afraid of, President Obama wanted more funding for technical education. Jobs of the future will be different, and we need to prepare for them.
Some areas where AI is already replacing humans
Just to keep the panic up for a little longer, let's take a look at some of the areas where machines are already proving themselves to be superior to people.
1. Strategy Games
The Chinese game of Go is a strategy game played on a board between two players. It is an ancient game that is intellectually challenging. Some call it one of the world's most complicated board games because of the sheer number of possible moves. Google's algorithm AlphaGo recently defeated current Chinese Go champion Ke Jie after playing against him a second time. According to Mr. Ke, the machine had been less humanlike in 2017, and more like a God of Go than when he played against it in 2016.
AlphaGo has been getting better at making unorthodox moves and challenging assumptions about this game. According to its creators at Google, the more AlphaGo plays (against itself), the better it gets. Rather like a human, but far more consistently. Mr. Ke has since said that he will give up trying to compete against AI, and stick to human opponents.
Self-driving cars are coming shortly. It will mean millions of drivers will lose their jobs. There are many benefits of a self-driving car. Driver error and negligence are responsible for most accidents, which won't be relevant when there is a precise machine driving unless there is a design or manufacturing fault. But professional drivers would have to find alternative means of livelihood.
AI can have significant scope in marketing. AI is any technology that mimics human intelligence, so it doesn't always have a robotic form. In marketing, chatbots can be AI. So can Voice Search, Dynamic Pricing, and AI generated content. Some content generation tasks AI will be particularly efficient in, include reporting sports matches, drawing up quarterly earning reports and compiling market data.
4. Military Technology
Around the world, AI and robots are changing the defense landscape. Here's an interesting example from this industry, which shows how job descriptions will change soon.
There is ongoing DARPA research about smart sensors that will be able to detect what is going on in battlefields, without running out of batteries and putting the lives of soldiers at risk. These smart sensors will run 24/7, and won't need a lot of power. They will stay dormant most of the time but still be sensitive to movements of troops or vehicles. This IoT (Internet of Things) device will also need AI to process noise and make sure that soldiers are receiving accurate information. AI in the military can help better protect human lives. But we'll need to train new types of engineers to interpret and design this data.
These are only a few examples in which we're already seeing signs of AI taking over human roles. In many of these examples, AI beats human intelligence.
Can AI Truly Replace Humans?
Can we say for certain that AI will replace human intelligence across the board? As Carl Sagan said, our having intelligence at all is a cosmic miracle. By creating an intelligence greater than our own through machine language, algorithms, and AI, are we setting ourselves up for a future where robots are self-aware and self-willed?
It's hard to tell. It depends on what we do with machine language in the future. Machine language runs on algorithms that we have designed, on a platform with an architecture that we've created.
For AI to truly beat human intelligence, it should be able to start with nothing but empty hardware and basic code commands (rather like a baby equipped with instincts that make it cry when it's hungry). It should be able to learn through experience. The AI should also be able to replicate itself or create another intelligence system on its own. So far, even the most cutting-edge AI systems are incapable of replicating itself. So, in theory, and practice, AI isn't close to matching human intelligence.
Experts say that the chances of human extinction due to AI is very slim - a mere 5 percent. And that may be because we're further ahead of AI, which is comforting to know.
What are your biggest fears about AI? What about your hopes for it? Tell us and let's get a fascinating conversation going.