Discussion Title: Will AI become conscious?

1. AI will become conscious one day.
1.1. Pro: Since there are robots designed with human-specific characteristics like [emotional intelligence](https://www.sapiens.org/technology/emotional-intelligence-robots/) that were previously thought not possible to have, consciousness may be another characteristic like that in the future.
1.1.1. Con: What looks like emotional intelligence in robots is a mechanically produced simulation, exactly like that of a doll. So just displaying it isn't going to be enough to create consciousness.
1.2. Pro: If AI can become more intelligent \(not just in certain areas, but altogether\) than humans, it will be smart enough to find a way to be conscious.
1.2.1. Con: We do not yet understand what function, if any, is served by consciousness, and an intelligent AI will not "find a way to be conscious" unless there is some value to it in doing so.
1.2.2. Pro: If AI builds different types of intelligence separately \(i.e. [specialization](https://www.albany.edu/sanestevan/pdf/T%20Hardy%20Yale%20honors%20thesis%20Maya%20Formative%20Ceramic%20Craft%20Specialization%20at%20San%20Estevan.pdf)\) enough that they could combine together, it can create [emergence](https://www.youtube.com/watch?v=16W7c0mb-rE), of which might appear consciousness.
1.2.3. Con: There are computers that can process at greater speeds than the entire human brain can, yet cannot be smart enough to think for themselves, as they don't think in the same way as humans. So skills are greater than capability in terms of being conscious. Without the skill, then being more intelligent alone won't work.
1.2.4. Con: Since intelligence could take place non-consciously \(as seen with [slime molds](https://www.youtube.com/watch?v=elqwn7k2Wwk)\), simply having it is not enough to lead to consciousness
1.3. Con: Right now, each AI is specific to one task. It'll be [really difficult](https://futurism.com/conscious-ai-decades-away) to combine all of them into one to mesh them together for them to work together.
1.4. Con: A lot of AI is [hyped](https://futurism.com/conscious-ai-decades-away). So adding in conciousness to AI, like the ones that exist in the movies, may not end up how we imagine it will be.
1.5. Pro: AIs are a reflection of our own intelligence. So if we are conscious, so can AI be as well.
1.5.1. Pro: Our thoughts and emotions in recent years have been found to be connected to a lot more than just our brains. We had to develop our own consciousness at some point in our history. In reality we are our own branch of AI.
1.5.2. Pro: The [processing abilities of AI](https://www.livescience.com/62656-when-will-ai-be-conscious.html) are not unlike those that take place in human brains. So if we keeping looking into how to replicate processing capabilities, then it may be possible with consciousness too one day.
1.5.2.1. Pro: [Sophisticated AI systems](https://www.livescience.com/62656-when-will-ai-be-conscious.html) use a process called deep learning, which is very similar to that of our own brains. It can be used to create consciousness, as that's what our brains use for it. Just like humans, to get from the ground level up to consciousness, time and more layers will need to develop. However, once it's at the human level, it'll possibly reach consciousness then.
1.5.2.1.1. Con: Deep learning is "inspired by" brain processes that use it for consciousness, yet we don't know enough about how consciousness works in the brain or where it comes from to know if it's possible or able to be recreated in AI, even with deep learning.
1.5.2.1.2. Con: There are indications that neural networks and deep learning may be [coming up to a wall](http://web.archive.org/web/20200221042716/https://www.wired.com/story/facebooks-ai-says-field-hit-wall/).
1.5.2.2. Pro: In a [neural network](https://www.livescience.com/62656-when-will-ai-be-conscious.html), deep learning enables AI to teach itself how to identify disease, win a strategy game against the best human player in the world, or write a pop song. If these seemingly impossible skills \(due to their need for creativity and emotion\) are possible to transfer to an AI, then surely consciousness will be possible to transfer too someday \(when the technology's there to process it\).
1.5.2.2.1. Con: AI may be able to teach itself, but it's not able to know when to use it or not without external input. Once AI knows when to use the tools their given \(like deep learning\) to decide and know when to apply it when fulfilling some need, then it's likely to become conscious then \(deep learning is not enough for this\).
1.5.2.3. Con: -> See 1.2.3.
1.5.3. Con: We don't really know what consciousness is in humans. Without this knowledge, it'll be difficult to add it to an AI.
1.5.4. Pro: If humans can develop an algorithm complex enough to match the human brain, consciousness could be developed
1.5.5. Con: Because it's a reflection and not a copy, it lacks the components that comprise our intelligence. A mirror showing a human may reflect their body but does not actually have one. The same could happen with AI: it may show, or simulate, consciousness, but not actually be conscious.
1.6. Con: It'll be difficult to get AI to truly replicate consciousness \(even with trying to mimic the human brain\).
1.6.1. Pro: We cannot make an AI conscious unless and until we know how consciousness was created in us. Consciousness in us seems to be some kind of an energy. Now, energy can neither be created nor be destroyed. But to make an AI conscious, we need to create a conscious energy which we cannot do if we do not know how that energy was created within us. If AI gets conscious without knowing how we became conscious, then it would most probably be by an accident or luck.
1.6.2. Pro: The concept of "consciousness" is not defined well enough for this to ever be properly evaluated
1.6.2.1. Con: Deep learning is the intelligence that is based on one definition of consciousness: [www.kialo.com](https://www.kialo.com/when-one-finds-what-works-for-them-that-they-dont-currently-have-and-doesnt-have-to-be-right-or-wrong-and-invests-in-it-31265.3?path=31265.0~31265.3)
1.7. Pro: The technology is advancing rapidly enough for this to happen.
1.7.1. Con: Even with the most advance technology, any [neural network](https://www.livescience.com/62656-when-will-ai-be-conscious.html) still relies on a human programmer setting the tasks and [selecting the data](https://www.livescience.com/58675-artificial-intelligence-learns-biases-from-human-language.html) for it to learn from.
1.7.1.1. Pro: [Consciousness for AI](https://www.livescience.com/62656-when-will-ai-be-conscious.html) would mean that neural networks could make those initial choices themselves. Since that is not possible with today's technology, consciousness in AI is not feasible.
1.7.2. Pro: A new type of consciousness will likely emerge and evolve around when singularity between man and machine appears: an AI far more advanced than today will ingest "consciousness" data \(likely from people\) until it becomes conscious in the process.
1.7.2.1. Pro: Due to humans merging with machines, humans will have greater capabilities than ever with consciousness: manipulating, expanding, understanding it, etc. These enhanced capabilities may make it possible to provide the knowledge and tools needed to form an AI consciousness.
1.7.2.2. Con: AI might just utilize our own consciousness when they combine with us for processing uses \(like a biological computer component\) rather than truly having their own artificially sourced one \(that may require being a separate entity from humans to truly emerge\).