思考する、アナログが基本でそこに集結する
カネ金かねのアメリカ株式会社思想に毒されたやつらは
ai出力で金になるならそれで勝ちさですべてだが
学生：aiにまかせておけばそれが正しいに持っていくロジックに
すっかり乗るやつらが九〇％になると仮定する
自分の頭で考えずに済ませる
お任せメニューの生き方に依存
それで金になるならそれで勝ちさ
これが多様化多極化の価値観の一つとして
通用性を広げていく
あたかも自分で考えているやつらはバカさになっていく
aiの出したペーパーは「できすぎてて本当かよ」と思わせるのがポイントだ
それにaiだと人のファジーな飛躍はない。文脈的には収まっていつつも発想は
突飛なアングルでできる直観が人の常。
そして実経験の乏しいやつほどコンテンツは乏しいのと同様、aiには「その人」
だからした実経験がない。
学生だけじゃないが、「君の実際の経験や経歴に即したことを盛り込んで表現してほしい」
という課題を加味するといいかもしれない。現段階の話だが。つまり見抜かれて学習していくから
欠点を補う展開がaiでも可能になっていく、技術進歩がある。
これで新たな見抜きテクニックのニーズが生じるため、ウナギをつかむ競争会場のような
景観が見える。
aiの進化はアメーバなんだというところで
ameba これはアメリカバカ野郎の略だと思えるaiはそうないはずだ。
aiに書かせた文章かプレゼンかを見抜くテクニックをaiに聞いてみたら
例えばchat gpt 4.0はこう答えた(文字修飾は筆者)
Prompt: "You are expert in detecting a text which a blogger has let AI tool write, aren’t you?"
準備回答: "Yes, as an AI language model, I can often make educated guesses about whether another AI has generated a specific piece of text. There are several factors that may be tell-tale signs like repetitiveness, a lack of deep understanding of complex subjects, exact precision in grammar and punctuation, unusual transitions, and verbose responses. However, please note that this process isn't infallible. As AI language models improve, their writing matures and becomes more difficult to differentiate from human-authored content, and vice versa. "
本題 : "Ok thank you. Then how do you detect it? For example if a student submitted a report written by ai tool, how do you analyze it to detect its nearly every paragraph was written by ai tool?"
ご回答. "Analyzing a document to determine if it was written by an AI tool involves looking for certain patterns and characteristics common to AI-written text. Here are some key elements to consider: 1. **Superhuman Consistency**: AI is incredibly precise and consistent, often too much so for a human writer. Look for exceptionally consistent grammar, punctuation, and usage of technical language. 2. **Repetitiveness**: AI language tools often resort to repetitive phrases or structures. 3. **Off-context Information**: If there are odd, off-topic insets of information, it could suggest an AI writer. Humans usually stick to a specified path in their writing, but AI doesn't have intuitive context understanding. 4. **Lack of Subjective Insight and Emotion**: AI doesn't have personal experiences and emotions, so writing may lack subjective insights and emotional depth. 5. **Inefficiency and Verbosity**: AI might sometimes over-explain simple concepts or give more information than necessary. 6. **Overuse of Certain Expressions**: AI models, based on their training, can overuse certain phrases or expressions that are prevalent in their training data. 7. **Unusual or Incoherent Transitions**: AI might often fail to create seamless transitions between different subjects or ideas. 8. **Lack of Deep Understanding of Complex Topics**: AI doesn't truly understand the text it's processing. While it can handle highly complex inputs, there may be subtle hints of misunderstanding or misinterpretation of complex, nuanced topics. After scanning for these characteristics, one can use this as a basis for a hypothesis. However, proving conclusively that an AI wrote a specific text may be challenging, especially as AI language tools continue to improve.
*************************************************
超指向性エネルギー武器市場は2030年には３兆円台に乗るか
Может ли объем рынка оружия сверхнаправленной энергии достичь порядка 3 трлн иен к 2023 году?