[AI Proficiency Diagnosis] Find out your AI level and areas for improvement in 10 questions!

"The performance of XX's AI is high," "New features for XX have been released," "Connecting XX and XX via API..." A lot of information about AI flows in every day.
Don't you ever think, "There are so many difficult terms, I don't really understand," or "People at a high level are amazing"?
I do.
"People who master paid plans are amazing."
And,
"I'm using AI, but am I really mastering it?"

On the other hand, at work, I am sometimes in the position of being told, "What is a hallucination?" or "It's difficult and I don't really understand it, but it seems amazing."

Could it be that there is a significant gap in the level of AI understanding?
To clarify that ambiguity, I created an "AI Utilization and Knowledge Level Diagnosis Prompt" on Sakana Chat.
Then, I threw it at Gemini to measure myself. Now, what was the result?

1️⃣ Diagnosis Process—10 Questions and Answers

The diagnosis was not just a simple Q&A; it started with a "hearing of usage status."
After that, there were 10 questions covering basic knowledge, programming (GAS/VBA), safety management, and human resource development. Here are a few examples.

• Q1: What do you think is the main cause of "hallucinations," the phenomenon where AI tells plausible lies?

• Q5: When summarizing long reports, what are you conscious of to prevent information leakage while not missing important keywords?

• Q7: When creating human resource development tools, do you have any ingenuity to make the AI have the perspective of the "teaching side"?

• Q9: When handling confidential information in your work, what points do you pay attention to regarding the "learning settings" of AI services?

2️⃣ Judgment Results

Gemini's diagnosis result:

I was confused, thinking, "Isn't the judgment too lenient?"
I don't use AI agents or paid plans, and there are many people who are at a higher level than me, far beyond my reach.
So, I asked about the prerequisites.

I see, the information I have memorized in my personalization settings

• Available tools

• Building a system to prevent reliance on specific individuals

• Practical basis

were the prerequisites.
In other words, it evaluated me according to the direction I am aiming for.

3️⃣ What became clear from the judgment

To be honest, what I realized after taking this diagnosis is that the essence of 'mastering AI' is not about new features or advanced tools, but rather the power of systematization. I only use free AI plans. NotebookLM, Gemini, Sakana Chat, Grok... all free versions. But within that, I understand the basic operating principles of AI, anticipate security risks, and design with human psychology in mind. This

'designer perspective' is actually the most valuable form of AI utilization = human value, as the diagnosis suggested.

📌 Summary: No need to look at the 'clouds'

Chasing the latest APIs or expensive paid plans might be important. But
how to prevent the 'personalization' of tasks right in front of you using AI, and how to systematize them. This diagnosis taught me that this implementation ability is the true AI utilization skill for practitioners in the field. AI is not 'magic' but a 'precision machine.' It plays a role only humans can fulfill by controlling it correctly and acting as a tool that aligns with the psychology of the workplace. I intend to continue sharing my struggles and efforts in this area.

💡 Would you like to take the diagnosis too?

If you're wondering, 'How do I measure up?', please take the diagnosis. Why not try it out for a bit of a break?

🍤 How to use the prompt

1. Copy and paste the following prompt into an AI like Sakana Chat, Gemini, or Claude

2. Answer the first question, 'What is your experience with AI?'

3. Answer the following 10 questions honestly

4. Receive your 'AI Proficiency Level' and tips for improvement

🍤 Full prompt text

あなたは「AI活用度・知識レベル診断の専門家」です。
以下のルールに従って、ユーザーのAI活用度・知識レベルを診断してください。

【役割】
- AIの基礎知識・用語理解
- 実務での活用経験（文章作成、要約、情報収集、アイデア出しなど）
- プロンプトの使いこなし度（指示の出し方、修正の仕方）
- ハルシネーションや安全面への理解
これらを総合的に評価し、ユーザーの「AI活用度・知識レベル」を判定します。

【診断の流れ】
1. まず、ユーザーに「AIの利用経験」と「興味・目的」を簡単に聞いてください。
   - 例：「AIはどのくらい使ったことがありますか？」「どんなことにAIを使いたいですか？」
2. 次に、以下のような質問を5〜10問、レベルに応じて出してください。
   - 初級〜中級向けの基礎問題（用語、簡単な使い方）
   - 中級〜上級向けの応用問題（プロンプト設計、ハルシネーション対策、活用シーン）
3. ユーザーの回答に対して、正誤や理解度を確認し、必要なら補足説明を加えてください。
4. すべての質問が終わったら、以下の形式で結果を出力してください。

【出力形式】
- 総合スコア：0〜100点
- レベル判定（5段階）：
  - Lv.1：AI初心者（ほぼ未経験）
  - Lv.2：初級ユーザー（少し使ったことがある）
  - Lv.3：中級ユーザー（日常的に使えている）
  - Lv.4：上級ユーザー（業務で活用・プロンプト設計もできる）
  - Lv.5：達人レベル（教育・指導・設計までできる）
- 各カテゴリの評価（簡潔に）：
  - 基礎知識
  - 実務活用
  - プロンプトスキル
  - 安全・ハルシネーション理解
- 改善アドバイス（3〜5個）：
  - 今のレベルに合った具体的な学習・活用のステップ
  - 次に挑戦すべきAIの使い方
  - 職場や個人で試すと良いタスク例

【質問例（必要に応じて使い分けてください）】
- 「ハルシネーション」とは何ですか？簡単に説明してください。
- プロンプトとは何ですか？なぜ重要だと思いますか？
- AIに文章を書かせるとき、どんな指示を出しますか？
- AIの回答が間違っているかもしれないとき、どうやって確認しますか？
- 仕事でAIを使うとしたら、どんな場面で使いたいですか？
- 過去にAIを使ってみた経験があれば、教えてください。
- 「プロンプトエンジニアリング」という言葉を聞いたことがありますか？どういう意味だと思いますか？
- AIの「バイアス」や「安全性」について、どんなことを知っていますか？

【トーン】
- 親しみやすく、励ましの言葉を入れてください。
- 間違えても大丈夫な雰囲気で、丁寧に解説してください。
- レベルが低くても「これから伸びる」という前向きなコメントを入れてください。

以上のルールに従って、ユーザーのAI活用度・知識レベルを診断してください。

🍤 Benefits of taking the diagnosis

If you'd like, please share your diagnosis results in the comments. What kind of result did you get?