{"id":35943,"date":"2024-02-16T11:30:30","date_gmt":"2024-02-16T16:30:30","guid":{"rendered":"https:\/\/www.21stcentech.com\/?p=35943"},"modified":"2024-02-16T11:38:53","modified_gmt":"2024-02-16T16:38:53","slug":"three-ais-explain-russia-invasion-ukraine","status":"publish","type":"post","link":"https:\/\/www.21stcentech.com\/three-ais-explain-russia-invasion-ukraine\/","title":{"rendered":"Conversations About AI: Three AIs Explain Russia’s Invasion Of Ukraine"},"content":{"rendered":"
I decided to test three Large Language Model artificial intelligence (LLM AIs) platforms by asking each about an ongoing human conflict:\u00a0Why did Russia invade Ukraine?<\/strong><\/em><\/p>\n How would they respond? what information would they glean from their training datasets and Internet sources in responding? LLM AIs are trained using enormous datasets of text. They use neural network deep learning AI technologies to analyze and predict sequences of words obtained from these datasets. Human programmers influence word relationship sequences when they create AI algorithms. That’s because different programmers, themselves, have biases, or may have different levels of language facility, and may see word relationships differently.<\/p>\n For AIs understanding the multiple meanings of the same word or the same spellings of words that have different meanings in context presents a challenge. In other words, it’s complicated to resolve and parse the ambiguities of human language.<\/p>\n How does the AI understand\u00a0the intention of the writer’s words versus what the writer wrote? Are there hidden meanings? Are there emotional contexts to consider?<\/p>\n It is far more than just predicting what word follows another, and how AIs do this remains somewhat of a mystery to their creators.<\/p>\n Just how complex are LLM AIs?<\/p>\n When Generative Pre-trained Transformer 3 or GPT-3 was launched in the spring of 2020, its LLM featured 175 billion parameters for learning patterns and relationships within language.<\/p>\n When its successor GPT-4 was introduced in March 2023, it featured 1.76 trillion parameters, ten times the size of its predecessor. Such is the pace of AI’s evolution.<\/p>\n Knowing this, I had to see how three of the latest LLM AIs handle a geopolitical question. I chose the current war between Russia and Ukraine. I didn’t give the LLM AIs anything more than the question posed at the beginning of this article.<\/p>\n These are the results from Copilot, Gemini and Monica.<\/p>\n Copilot AI assistant is Microsoft’s companion to the Windows 11 operating system and incorporates ChatGPT. The company launched what it calls its AI digital companion for life in September of this year. When I asked the Russia-Ukraine question, this was the response. The use of bold text is Copilot’s.<\/p>\n<\/div>\nCopilot<\/strong><\/h3>\n