September 2023

65 64 Warren McCulloch and Walter Pitts defined the first mathematical computational model similar to a neural network in 1943. They created the first artificial neuron model. Since then others have built on their work to create the neural network systems that we now know today. When you ask AI like ChatGPT a question, it does a good job of providing a response that is convincingly human. It is easy to believe it understands what it is saying, but it is using probability to predict a response. So remember, that at the moment, AI can’t think or feel. When chat bots make things up, it is known as an AI Hallucination. It generates rather plausible replies but cannot assess the truthfulness of its responses. So, anyone using generative AI needs to remember that these mistakes can be spotted by someone who knows the topic . If AI is trained on data that is racist, biased, or hateful, then it’s output will be too. For example in 2016 Microsof t launched an experimental chat bot Tay. But Tay was quickly shut down af ter making racist and of fensive comments af ter interacting with users. In amongst all the caution and calls for improved safety features, it is true that AI has many potential benefits. It can enhance and speed up work in many areas such as healthcare or programming. It is up to the users and creators to work together to decide how and when it should be used, if at all. AI isn’t as new as you may think AI cannot feel or think AI often makes things up AI can be biased, racist and sexist AI has many strong benefits. Adapted from the BBC/Open University video “Five things you really need to know about AI” Human interactions are intricate, with layers of hidden thoughts and emotions bubbling beneath the surface. Building friendships requires time and trust to uncover the genuine person beneath the facade. Understanding others’ intentions is challenging, as we can’t know what’s in their minds. Given these emotional complexities, wouldn’t it just be easier to rely on AI as a trusted friend? It is always available, it won’t share your secrets, and it will know something about every interest you have. According to Aristotle, there are three kinds of friends: useful friends, friends, with shared interests, and true friends. Emotional connections are foundational in each. Useful Friends Useful friends help you because they want you to do well. AI is incredibly useful across multiple facets, but unlike a friend, it has no intentions behind its help. It’s programmed to analyze existing data, and its responses are devoid of personal investment and emotion. While it may seem that useful friends don’t require emotional connections, AI is fundamentally different. It assists because that’s its job. AI cannot have interests or emotions, preventing it from forming emotional bonds. Regardless of the type of friend, they should want to help you because they care about you. Friends with shared interests Friends with common interests must have emotions to have interests. AI can engage in conversations about our interests, but it cannot share or exchange interests on an emotional level. For example, if you enjoy songwriting, AI provides information on the topic, and may create music, but it won’t share your passion or joy of the creative process. True friends Aristotle said true friends are alike in virtue and wish each other well. True friends are useful and share interests. You each have a personal investment in the relationship. I care about my friend’s lives and want them to be happy and prosperous. I can’t share jokes with AI hoping for a laugh or confess my dark secrets and expect understanding. Aristotle is still relevant So, it seems that Aristotle’s thoughts are still valid. It is impossible to be friends with AI. It can’t feel emotions, and I can’t connect with it emotionally. The complexity of our emotions is what makes human relationships beautiful. It can be challenging to open up to strangers, but when we do, it can lead to deep and meaningful relationships, a unique aspect of being human. Discovering someone who cares about us, understands us, and shares our interests is exhilarating. Such connections are built on shared experiences and trust and offer a sense of belonging and connection. So, although AI may offer an uncomplicated presence, I don’t see it will ever be possible to hope of real friendship with AI. By Victoria Deckard, Grade 9, Bloomington South High School, Indiana Friendship 2.0: Is AI A Better Kind Of Friend?