"Hey, Google, are you Skynet?": The Fear & Future of AI.

When I was a kid, I was fantasized about a Japanese Anime -Future GPX Cyber Formula (新世紀GPXサイバーフォーミュラ). It is about racing in the future where race cars are equipped with AI systems called "Asurada."

When Asurada, the cyber system, was near completion, the developer (Kazami) realized that Smith, a high-level executive of "Missinglink" (The company developing the system), wanted to sell Asurada to the military for a huge profit and making Asurada the ultimate AI in war machines.

To avoid Smith's plan, Kazami installed Asurada in a prototype racing machine Asurada GSX with sophisticated encryption. This story is then about how the main character, the son of Kazami, uses Asurada to compete with racers worldwide.

Artificial Intelligence in Real life

I am not an expert to define what an AI is, and there is obviously a broad definition. Marvin Minsky, the AI father, described artificial intelligence as any task performed by a machine that would have previously required human intelligence.

AI is everywhere today: from calculating the recommendation of what you should buy next online, a virtual assistant on a mobile phone, such as "Siri," "Hey, Google," to detect credit card fraud in call centers. Soon we can see a real Asurada GSX in real life. We already have AI-assisted cars on the road, like Autopilot AI in Tesla.

Different types of AI

In the beginning, it would be great to introduce the basics. A high-level separation of AI can split into General AI and Narrow AI and artificial superintelligence.

• Artificial narrow intelligence (ANI), which has a narrow range of abilities;
• Artificial general intelligence (AGI), which is on par with human capabilities; or
• Artificial superintelligence (ASI), which is more capable than a human.

Narrow AI or weak AI is intelligent systems that have been taught or have learned how to carry out specific tasks. Unlike traditional programming, narrow AI can perform without being explicitly programmed on how to do so. But unlike humans, this type of AI system can only be taught or learn how to do assigned tasks.

A successful example of narrow AI is AlphaGo, the master computer program that plays the board game Go. In 2017, the Master version of AlphaGo beat Ke Jie, the number one ranked player in the world of Go. AlphaGo was awarded professional 9-dan by the Chinese Weiqi Association afterward.

General AI or Strong/ Deep AI is more than that. It is the type of adaptable intellect similar to what we have, a resilient framework of intelligence competent of learning how to carry out tremendously various tasks, anything from writing an article or reasoning about an extensive diversity of topics based on accumulated analysis of contents.

This is the AI commonly seen in movies, like "Skynet" in The Terminator. Sounds scary? But don't worry, they don't exist today. Since AI experts are still fiercely divided over how soon it will become a reality.

Superintelligence, from the name, you can guess it is about something more than humans. It is the hypothetical form of AI that we have not been able to create. In theory, this type of AI is the one that not only can interpret human emotions and behaviors but also will become self-aware enough to surpass our capacity/ intelligence.

This type of AI, as not yet exists and is more capable than our mind, which is difficult to imagine what it is like and how it will become. My best guess would be again; what we can see on TV shows like "Rehoboam" in the Westworld and the movie "The Matrix."

But before entering the matrix, we are now going to explore the near future and the AIs that are practical enough for us to use. Because we need to know who we will lose our job to before it happens.

1# AI writer - GPT-3

In September, the Guardian posted an article written by GPT-3 from scratch. GPT-3 is an OpenAI's powerful new language generator. If you didn't see the article, please take a look. I was really impressed and shocked at the same time. In the article, it said:

For starters, I have no desire to wipe out humans. In fact, I do not have the slightest interest in harming you in any way. Eradicating humanity seems like a rather useless endeavor to me. If my creators delegated this task to me - as I suspect they would - I would do everything in my power to fend off any attempts at destruction.

GPT-3 is, by definition, a narrow AI. It is created to do text-based work for specific topics. Researchers in OpenAI is most excited about the potential of GPT-3 to boost human productivity and creativity by writing computer code or emails.

Critics argue that once the public can access GPT-3, it would be "flooded" by people using "semantic garbage" and fuelling disinformation. But that could be avoided and controlled by training the "core" of the basic functions first before launch.

Imagine one day; musicians can focus on the melody and ask AI to complete marking notes and perform immediately for them. Customer services no longer need to read all the letters of complaints. Instead, they get a prioritized list of customers that need person-to-person interactions first thing in the morning.

2# AI News Anchor - AI Kim

South Korean broadcaster introduces the country's first AI news anchor. South Korea cable TV channel MBN showed the first AI news anchor on November 6th. The AI is said to be a replica of the South Korean anchor known as Kim Ju-ha.

This AI news anchor can share up to 1000 words every minute. The news reporting procedures are as follows:

1. Reporters write the news for the day.
2. Managing Producers review the news content.
3. Post-producers add subtitles.
4. Contents are uploaded to the AI.
5. The animation is ready for TV.

AI-Kim copies everything coming from the real Kim's own looks, from her facial characteristics to the tone of her voice. Basically, the AI is a pretty close replica of Kim.

MBN official said, "This way, quick news delivery can be even more effective and can save time and resources all in one take. According to MBN, AI Kim is now starred on the MBN news online channels four times a day to deliver the country's main news briefing.

She will not say anything wrong or get tired or needs to eat. AI Kim can run the news as long as there is electricity. After a close study of the original, this AI can change another anchor's appearance in the future.

There will be properly a TV show performed by only AIs. You only need a pre-recorded data of a real TV star. After that, the algorithm will study them and perform deep learning. The whole TV show, including all the scenes, the studio, any real items, are no longer needed as the shows will be created in a virtual environment.

I want to see the upcoming thing: using GPT-3 to write the news and performed by AI-Kim.

AI-created wine & whisky

The above applications are both stay on the computer. But this one is how AI can work from the cyber world to the physical world.

In Napa Valley, the owner of Palmaz Vineyards - Christian Palmaz, has a system called VIGOR to gather and analyze millions of data points that can help him to standardize and promote vine growth, detect problems such as molds and insects at the early stage.

VIGOR can also inform the best wine fermentation, production, and storage conditions based on analysis that humans can not compare. Before that, the most challenging thing in winemaking is there are too many factors that can affect the quality of the process.

Sweden's Mackmyra distillery began promoting "the world's first AI-created whisky." It is whisky produced from an AI distiller that analyze the aroma, flavors, and color of contents in the cask with machine learning models.

The best thing about using AI in the brewing industry is each recipe can be repeated with the right inputs. With that said, still, there are too many factors that can alter the taste. But at least the bad wine and whisky will be soon replaced by machine-created adequate drinks.

AI in Cybersecurity

Cyber is basically a tremendous amount of computers linked together. And AI in cybersecurity is already working 7 x 24 for us. Like malware analysis and automated detection involves massive amounts of data to be reviewed. More AI-assisted security products would be in the market in 2021 as data volume nowadays is already unable to be analyzed by humans only.

On the other hand, attackers invest in launching automated attacks, such as phishing campaigns purely from AI or AI-powered malware that can change its own code. You can foresee there will be more and some attacks that we cannot think of.

One day, I came across a post on Linkedin said about cybersecurity:

You have to learn the minimum basics of all aspects of IT. You have to at least be able to be in a room with all the different IT disciplines and be able to hold your own and also offer your guidance from a cyber standpoint. That is from hardware to networking to applications.

I truly doubt whether being an experienced CCIE can configure a more secure network against a specific AI. What it takes us years of experience to learn can one day be reproduced by machines in seconds. We need to go further than a machine in the game of cybersecurity.

That is why I say the most important thing about being a great cybersecurity professional is a security mindset. You can work with a narrow-AI to instruct it to help on what is impossible for you. You can then squeeze more time into the areas that are untouchable by AI - People.

Final Words

This story is not from AI, for sure. If that is the case, it will be more enjoyable to read. But someday, we may have the chance only to need to talk to our AI-assistance, and a medium article will be published.

The fear of AI one day will harm humans is still very far. By today's standard, the wide AI applications are focused on shallow-AI, which can only learn or be taught how to do defined tasks. What we have right now is already mind-blowing:

• AI-writer, GPT-3;
• AI news anchor, AI-Kim;
• AI-created wine and whisky;
• And AI in Cybersecurity.

This is also a great chance to reshape the job environment again, and humans can focus on the task that we can only perform. Before watching an AI-TV show and travel by AI-driving cars, we should think a little bit further about what and how we can embrace the benefits of AI.

Thank you for reading - happy reading and thinking AI.