The generative adversarial network demonstrates how far humans have come in regards to technology over the last centuries. It is easy to assume that the images generated by the A.I. are completely on its own merits; however, humans are ultimately the brains of the operation. After all, facial-recognition algorithms had to be developed by people before the machinery could function.
Technology is only as good as the research and skill that goes into its development. This is highlighted in the article, in which the author describes some of the flaws in the A.I. system. For example, it has trouble registering accessories, such as jewelry and glasses. Additionally, the system exclusively generates faces that are completely symmetrical, which is not the case for real human faces. The author poses an interesting question: “Do we place too little value in human intelligence — or do we overrate it, assuming we are so smart that we can create things smarter still?”
Personally, I argue that we underestimate human intelligence in comparison to artificial intelligence. Technology depends on the input of human efficiency and their own intelligence. As self-sufficient as machines may seem, there is always room for error because they reflect human errors. Furthermore, people often fear the potential and dangers of A.I., when the real threat is the people themselves. Individuals could use the A.I. technology to create fake social media accounts and scam other users. We should be blaming the technology, however, we should blame the people capable of committing such acts of fraud.
The biggest challenge with AI is its need for large amounts of data to properly function. We give AI this data when we publicly upload things to the web. The process of grabbing publicly available information is called data scrapping and it’s the easiest and most readily accessible way for AI to obtain information. From this pool of information, AI technology extracts generic data that is readily accessible to it and rearranges it. The technology doesn’t analyze or interpret the information in meaningful ways it just sort of collects, stores, and repurposes it. Dating mining or scraping however is involved in extracting value or meaning behind the data. It’s for this reason people have established careers in cybersecurity.
I agree with you that AI is not bad. It does get the blame though when people use AI data to accomplish dishonest or unethical acts.
This was so interesting to read! I agree that humans really the ones behind AI, leading these machines to sometimes make human-like errors. AI came from a person, and that person is the one who should be held accountable.
It does worry me that we call these complex computing systems intelligence; are we pulling computers up to our level, or are we reducing ourselves to machines? It seems even weirder when you think that these systems are only as powerful or as “intelligent” as we are; AI reflects our own biases, but also can’t reflect on these biases. I don’t think we should blame the technology; we should blame the people who use and abuse it.
Interesting post! I never really thought about how AI is only as smart as the humans that create it. I had no idea that people could buy “fake” AI-generated faces to use on social media platforms and such, and I agree that we shouldn’t blame the AI for these issues but the people themselves that choose to be that dishonest and sketchy.
One of the things that is strange about this moment of AI is that the use of a generative adversarial network largely means that the computer is teaching itself what a “face” is (or how to play “chess” or “Go”). Of course, someone had to program the basic of the GAN, but after that point, computers can do a lot of the work all on their own.
Love this post! It reminds me of the argument that A.I. is just a tool that can be used for good or bad, same as social media, weapons, all those standard sticking points. Some things are morally or ethically neutral, and when it tips to one side or the other, it is usually the work of a human, not a machine.