Artificial Intelligence (AI) is a form of computer which has the ability to think for itself and perform tasks that usually require the intelligence of a human such as: speech recognition, decision-making, visual perception and translation between languages. However, there is a common misunderstanding that there is just one type of artificial intelligence. This is not the case. There are actually 3 different types of artificial intelligence: ANI, AGI, and ASI. ANI, or Artificial Narrow Intelligence is the weakest kind, hence it sometimes being called Weak AI. This type of AI can only specialise in one area such as beating someone at chess, however it cannot do any more than that. AGI, or Artificial General Intelligence is the next stage up to ANI. This form of AI is a computer that has the same intelligence as a human and so it is sometimes referred to as Strong AI. AGI can usually carry out tasks that people would generally do, like plan and solve problems. Finally ASI, or Artificial Superintelligence, is the type of AI that could potentially cause the extinction of the human race. This type of AI has capabilities that are above those of humans as its intelligence is higher than humans .
Based on the definitions from waitbutwhy.com ANI is harmless and cannot possibly be a threat to humanity. In the United States of America in 2017, over 85 billion people have access to a form or ANI also known as Siri . Siri has many functions and can perform basic tasks like checking the weather and sending texts to family members. In times of emergency it can even save lives and call the emergency services which in turn protects humanity and life. So how can it be a large threat to humanity if it is saving lives? The source that was used to find these statistics started out on Twitter, a widely used social media site. The statistics which are put onto this website are managed and edited by Craig Smith meaning that they are not checked by a third party and so they could be false statistics and so could be inaccurate. This could mean that the website is an unreliable source of information. Other forms of ANI include: self-driving cars, Cortana, Alexa, AlphaGo, and Google Assistant. Many of these forms are used to help us in every-day life to do basic tasks such as setting reminders to go to the shop or to quickly search the internet for research purposes. Self-driving cars will change how people live in significant ways. The boss of General Motors (GM), Mary Barra, describes how self-driving cars will cause “zero crashes, zero emission and zero congestion” which would in turn reduce travelling times, save lives and prevent injuries, and would also help to prevent climate change. At the moment, on average 1.25 million people die from car accidents and an extra 20 million are injured each year around the world. Self-driving cars can cut this down to 0 people being killed and fewer people being injured . This is clearly a positive effect and therefore shows that the threat of AI to humanity is negligible. The source used for this part of the argument is an English newspaper that is edited in London. The newspaper is not very biased at all but does tend to be slightly biased to the right-wing and so some of the information in this newspaper could be seen as biased. With it being a newspaper, all of the facts are back-checked and are very likely to be accurate and there is a low risk of them being completely false and it is therefore a reliable source of information.
Despite how much AI is currently helping us, we are only using ANI which is the most harmless form of AI as the definition from waitbutwhy.com said. ANI has very little intelligence and therefore cannot think entirely for themselves meaning that it cannot see human flaws and so has no reason to not help us. However, AGI and ASI both have intelligence at human level and above it. This could result in them seeing human flaws and so may decide to abandon helping our species and could potentially start trying to take over. This is currently only believed to be science fiction however it could occur in extreme circumstances. If we manage to create ASI then we don’t know that it will still help us, and we are gambling with something we cannot comprehend. We have already seen that AGI has the capability of creating new things for example, a new language. In 2017 social media company Facebook created Artificial Intelligence with capabilities that were closer to our own than basic AI. The Artificial Intelligence, in the form of a chatbot, that they created eventually created its own language and used that language to communicate with another AI. This would have been perfectly fine if the scientists were able to decipher the language that they had made but this was not the case and so the scientists had to shut the program down. This event came after scientist Stephen Hawking to make a statement saying that AI “would take off on its own and re-design itself at an ever-increasing rate” before going on to say: “Humans who are limited by slow biological evolution, couldn’t compete, and would be superseded”. Other tech luminaries, for example Elon Musk and Bill Gates, also joined Stephen Hawking by warning about the unforeseen consequences that AI could possibly have on the human race . This could suggest that the production of Artificial Intelligence with the capability to “redesign itself” is a very large threat to humanity and is one that is rapidly approaching and should therefore be looked into to minimise the risks of the end of humanity occurring. The opinions of Stephen Hawking and the other tech luminaries are most likely to be biased. They have formed a strong opinion on the issue of AI and therefore, since this is simply an opinion, the overall factual content could be completely inaccurate and should be viewed with some caution.
The article by Forbes suggested that there was widespread panic within the laboratory creating the Artificial Intelligence which deviated from English, however this information is not completely accurate. The lead author of the study from Facebook, Michael Lewis, said that “there was no panic and the project hasn’t been shut down.” Later in the statement, he said that Facebook “used some additional techniques to get the bots to work as we wanted” which clearly shows that the Facebook technicians who were working on the chatbots are able to control the form of AI and could change the programming to prevent the AI from going against the purpose it was created for. This consequently lowered the risk of the AI becoming a threat to the company. This idea can be used on a larger scale to prevent any form of ASI being created. This will limit the overall risk of AI being a threat to humanity and therefore because programmers will be able to control the AI. This will prevent the creation of ASI and therefore the threat to humanity is very low, almost negligible. The statement made within this article was from a technician working at Facebook on the chatbot experiment which could mean that the statement was made to cover up the truth behind the experiment. The technician could have been trying to hide what might have been a failed experiment and this source could therefore be completely unreliable. Many other magazines and websites, such as cnbc.com, also agree with the article made by Snopes and said that certain media reports were “clickbaity and irresponsible”. The article then mentioned that the chatbots made by Facebook were “simply a research experiment” . The website say that they got this specific information from Dhruv Batra from Facebook’s AI Research group meaning that this information is also directly from Facebook. Therefore, just like the statement made by Michael Lewis, the information could be unreliable as it could be being used to cover up the truth behind the experiment.