By Rishi Sharma
Humans have been fascinated by Artificial Intelligence (AI) for quite some time now. While a lot of us adore listening to Siri’s bad jokes, some of us have serious concerns about the consequences of the suddenly growing AI industry.
In the past few years, humans have invented numerous types of devices that feature the use of autonomous technology. Though the majority of us are not heavily interacting with or heavily relying on AI today, it is predicted that this is subject to change. About a decade ago, AI was man’s greatest fantasy; today, we have it in our hands, within our smartphones. Give the timeline, another decade, and we will probably be interacting with AI in our day-to-day life. With the sudden increase in technological advancements, it is worthwhile to weigh the advantages and disadvantages that come along with it.
A major disadvantage of almost every piece of technology is that it lacks perfection. If AI was to really replace humans in various positions – as drivers, soldiers, doctors, etc – they require constant maintenance that prevents hardware degradation. Most of us have already seen examples of such degradation in existing technology in our phones, laptops, Smart TVs, etc. A brand new iPhone is nothing quite like a glitchy 2-year-old-iPhone, and that’s a problem because it has a lot to say about how AI, being yet another creation of technology, might crash at any moment. Giving consideration to self-driving cars, crashes in the AI system may result in serious injury or death.
At this year’s Consumer Electronics Show, LG showcased a home assistant robot called CLOi. The presentation was going as planned until suddenly, the robot began “disobeying” the commander by ignoring him. It is crucial to note that even a small AI code produced by big tech firms such as LG is subject to crashes. In a bigger AI code, the probability of encountering a similar, or possibly worse, crash is much higher. The video of the exposition of CLOi can be found here.
A glitch in the code can cause the device to perform a function it was not meant to perform, and not much control will be given to humans as the device is autonomous. This can be rather disastrous in autonomous weaponry, which will compromise national security and the safety of soldiers. A reason we may not really see this as a possibility today is because AI is not used in the current-day battlefront very frequently, but that is highly likely to change in the coming years.
But that doesn’t mean the production of AI systems have to be stopped, right? As Mohnish Pabrai, an American businessman, says, “Mistakes are the best teachers; one does not learn from success”. Humans incorporate this saying in order to make the updated product better than ever. The discovery of a glitch eventually leads to developing ways to abolish that glitch from the technology. If a mistake never occurs, technology can never advance. In one of the United States’ first rocket launches, a NASA’s space shuttle, the Challenger, there was a malfunctioning disintegration system, which resulted in the death of all 7 crew members. But if the Challenger had never malfunctioned, NASA wouldn’t have ever been able to create the space shuttle that allowed Neil Armstrong to step on the moon. Technology, indeed, is progressive. A glitch in AI today could improve the entirety of AI technology tomorrow.
However, there is an important element of these systems that may be critically problematic. It is no doubt that the modern world has created numerous cyber criminals capable of hacking into even the most advanced encryptions. AI, here, is vulnerable to hacks. Although AI does not have to always be connected to the internet, it still runs a code; as long as there’s a code, it can be hacked. Alternatively, some AIs do need require a connection to internet. Self-driving cars, which are likely to get heavily prevalent in the future, will need to regularly update any glitches found within the code. Obviously, it is impossible to reach out to every single car every time there is a necessary update, therefore leaving these cars highly demanding of an internet connection. Self-driving cars produced by Tesla Motors already feature an example of this connection. The same can be said about autonomous weapons, as the relevant authorities have to constantly update the target list. Requiring an internet connection makes self-driving cars and autonomous weapons highly vulnerable to remote hacks by terrorists.
As reported by MIT Technology Review, General Motors CEO, Mary Barra, made a public announcement regarding this issue. “The threat landscape is continually evolving, and sophisticated attacks are specifically designed to circumvent even the most robust defense systems,” she said. “Whether it is phishing or spyware, malware or ransomware, the attacks are getting more and more sophisticated every day.”
In the case of autonomous taxis, having the car’s AI offline is possible. All taxi drivers are required to leave their cars at certain locations once their job is done for the day. At these locations, the updates can be installed in many cars, collectively. Since it is only the AI system that is offline, there will be no disruptencies with the taxi’s GPS and other non-AI components that require an internet connection. Although containing the code for the AI, these taxis have a hard-case that engulfs the machinery that runs the code, giving it an anti-theft feature, or as said in technological terminology – a physical encryption. There is definitely a great place in the future for autonomous taxis.
The same can be said about most public transport services. In bus services, a similar system as found in the taxis can be implemented. Since all busses have to be returned to a designated outposts (mostly bus stations), it is also very easy to insert updates collectively.
AI will be changing our lives. Many people, however, are worried that this change may be for the worse. This worry engulfs the fact that people believe that their jobs will be replaced by robots. In a survey conducted by the Pew Research Center, 73% of participants are worried or somewhat worried regarding AI taking replacing humans in the workplace.
Source: Pew Research Center
If this were to occur, the demand for products featuring AI would also increase, thus creating many other new jobs. However, such new jobs would also require great academic conditioning. In truth, AI is likely to replace jobs such as cashiering, janitoring, delivering goods, etc. These jobs are usually occupied by people that have not met the academic criteria for other jobs. These people will not be accepted into the new jobs that have been created due to their academic inabilities, and based off this, it can be hypothesised that unemployment in such people will increase significantly.
Conversely, in medical jobs, the use of AI may be beneficial. Diagnosing patients will be more efficient and much faster than if it were to be done by a human. Even in developed countries, such as the United Kingdom, many patients have to wait hours or sometimes even days to be diagnosed and treated (as reported by The New York Times). This problem has arisen due to the fact that medical work requires a lot of manpower; moreover, all employees need to satisfy a set of very challenging academic requirements. To add on, humans, under the pressure of time, are subject to making numerous mistakes. All of these problems can be solved by AI in hospitals. The fact that these robots will have a computerized intelligence of their own will allow them to calculate risks and immediately make decisions based off that, performing their job very quickly. Timing is very important in patients with fragile health. From drawing blood to performing sharp surgeries, a robot with AI can work in various medical fields alongside nurses and doctors and improve the efficiency of a medical workplace.
Technology has become so integrated into our lives that we have placed heavy dependence on it. Artificial Intelligence, a feat of technology, is bound to change our lives. AI can have different impacts depending on where and when it is used. By allocating it in the right places at the right time, AI is undeniably a step forward for our society. It truly comes down to how, where, and when we plan on using AI.
“The first rule of any technology used in a business is that automation applied to an efficient operation will magnify the efficiency. The second is that automation applied to an inefficient operation will magnify the inefficiency.”
~ Bill Gates, co-founder of Microsoft