Why Artificial Super-Intelligence Concerns Me
Artificial Intelligence Does Not Concern Me, but Artificial Super-Intelligence Frightens Me
However, suppose the capabilities of these lovely robots fall into the hands of people who have no concern about human suffering and even enjoy killing them for their enjoyment. If that were to happen using artificial super-intelligence, the apocalyptic future imagined by The Terminator movies could become a reality.
This article is not about giving answers or scaring people but to ask some informed questions to create awareness of the need for society to take a long and hard look at the ethical issues surrounding artificial super-intelligence and other advanced technologies supporting this movement.
I do not wish to discourage the progress of technology. On the contrary, I am a technology advocate and have spent more than four decades of my life studying and working on the technology of artificial intelligence.
Over a decade ago, when I mentioned the risks of social media in development communities, some people called me paranoid, and some found my caring ideas for user community fictitious. Now we see social media tools can easily change and favor the creators of algorithms or those who sponsor them.
For example, one of the social media tools was designed to create a family atmosphere. We all embraced it, and it grew so quickly. Now, it transformed into a tool that millions of people use for different purposes.
We cannot underestimate the power of intelligent machines. When they are programmed well with the power of machine learning and neural networks, they can surpass our abilities in certain areas.
Who wasn't surprised when Blue Gene beat chess champion Garry Kasparov in 1996? Even ten years ago, Watson beat the two greatest Jeopardy champions, Ken Jennings and Brad Rutter. These are just two examples from a single organization, IBM. There are thousands of business organizations and research institutes creating similar capabilities.
Is anyone controlling those proliferating technology constructs? Who can check trillions of lines of codes written in hundreds of different languages? And who can control the codes written by the sophisticated artificial intelligence systems? Once AI systems start working in concert as sophisticated GANs (generative adversarial networks), the advancement outpaces our ability to foresee what needs to be regulated for the public good. Deep Fake techs and their misuse are chilling examples.
In the year 2003, I was a consultant for an organization in the financial and banking industry. It was one of the largest banks in the region. Security is of the utmost importance in this industry. So, it was not difficult for me to justify the security budget that I created in the bill of materials in the security architecture I proposed.
The solution was successfully implemented. In a specific branch of the bank, in addition to many security measures in industry compliance, there were 17 firewalls providing access controls. However, some business people characterized the solution as overkill.
I understood their concerns and explained to them the complexity of online banking. An executive even mildly insulted me, saying I was exaggerating the power of hackers. Then, one day at 2 AM, when I was in my deep sleep, I had an emergency call asking me to immediately go to the venue to support a team of 300 engineers in a critical situation. The bank's Internet service branch was hacked, and the case was disastrous, causing the bank millions of dollars lost every hour.
I share this story because, based on the knowledge at the time, we created a paragon of architecture, spending millions of dollars on developing a secure system. Despite its architecture, design, implementation, and stringent operations, hackers found a way to break the system. It was a wake-up call for many other organizations that underestimated the value of cybersecurity investments at the time.
This is only one example. I witnessed many more situations over the last two decades when customers joined the Internet. Some naively believed that by just adding VPNs (Virtual Private Networks) and firewalls, everything would be fine. But cybersecurity is so complex. I worked with remarkable ethical hackers who explored unique ways hackers created. Some were so out of the box and even beyond the galaxy thinking.
With so much knowledge and power in cybersecurity, we still struggle. Many prominent business and government organizations are hacked and with alarming frequency. I don't have to convince you about these incidents as they are in the media as public news. But, of course, some of them do not go public for various reasons.
Adding measures and increasing our capability is not enough to control the Internet even though it is still in its infancy. The Internet of Things (IoT) is growing so rapidly. Billions of devices are being added to the IoT ecosystem. Now, some are working on the Internet of Bodies, including human intelligence to IoT networks via embedded body sensors.
Some technology professionals and leaders with good intentions see the concerns as exaggerated. For example, some think that the technology we have developed so far is primitive that creating an artificial super-intelligence would be impossible in our lifetime.
I love the optimism and support my optimistic friends as they may live longer than others. Some of those great people think that we have developed great technologies like blockchain, creating trust. I love blockchain. I have massive respect for Satoshi Nakamoto and his brilliance. However, staying in the reality zone and with all due respect to his intelligence, I also believe there will always be more intelligent people on this earth. Until 2009, no one knew about the intelligence of Satoshi, but he appeared out of nowhere and shook the world
More interestingly, some even asked me to show evidence from the scientific literature whether machines have ever overtaken humanity in history. I respect their intellect and a false sense of confidence but cannot help myself smile when they challenge my concerns with unscientific arguments like this.
There are, of course, some naïve technicians who tell me: "Stop worrying mate, we can turn off the damn things if they misbehave! They have an on and off switch." They are good friends, so I won't be rude enough to use the analogy of emus putting their heads into sands. Instead, I just share my recent article about what happened to the deep fake videos.
We think we know a lot showing our body of knowledge. In reality, our knowledge is so tiny. However, our ignorance is endless. I focus on my ignorance to stay curious than hiding behind the tiny body of knowledge I have accumulated over the years. All we know as humanity is not even a drop in a vast ocean.
We cannot even compare known with unknowns, let alone unknown unknowns. However, we know that there are millions of people who can think beyond the box. So our technical and scientific knowledge cannot protect us from unknown risks.
Currently, we are capable of controlling artificial intelligence, and our lovely robots like Sophia are mainly empowered by cognitive computing, which has almost become a simple technology in many tech communities. Anyone can access very sophisticated codes and create open-source ones which cannot be monitored and go viral. We cannot even control technology shocks disrupting global economies.
Besides all technology risks and issues, I have a serious concern about artificial super-intelligence, particularly when this super-intelligence is developed and run by people with wrong motives.
Quantum computing is not too far. There are already several companies that created quantum computers. Many tech communities are working hard to create quantum code and hardware. We cannot underestimate the power of quantum computing, especially when artificial intelligence platforms start using it as a major tool to grow.
Unless we can develop artificial intelligence systems with genuine empathy and compassion for humanity - not merely cognitive empathy as depicted by Sophia - super-intelligent AI systems could undoubtedly create derived values for their own survival mechanisms more than that of humanity.
Thank you for reading my perspectives.
The original version of this article is published on another platform.
If you enjoy writing, you can join Vocal as a creator to find your voice and reach out to a broad audience.