引導式測驗 · 中等 等級
警察社群媒體監控測驗
測試警察社群媒體監控:公共資料限制、帳戶驗證、背景、人工智慧分類、政策控制、保留和可審計性。
問題 1 的 8
A classifier labels a public post as threatening. What has the system established?
預覽每個問題8 問題
當您參加測驗時,就會出現答案和解釋。
- A classifier labels a public post as threatening. What has the system established?
- That the author intended violence
- That the account belongs to the person under investigation
- Only that the post matched its learned or configured criteria and needs contextual review
- That police may search the person’s devices
- Why preserve replies surrounding a post selected by a keyword search?
- The surrounding conversation may change the meaning of the isolated text
- Replies determine the platform’s retention policy
- Context always proves the post is harmless
- A keyword search cannot record timestamps
- Which policy issue does BJA’s social-media guidance specifically address?
- A universal federal ban on public-post searches
- A single model’s classification threshold
- Mandatory retention of every public post
- Access, use, storage, and dissemination of information obtained from social sites
- An entity-resolution tool links two accounts with similar names. What should the analyst do?
- Merge the accounts because the names match
- Verify identity with reliable context before attributing either account to a person
- Treat the highest-scoring profile as confirmed
- Delete the query log
- Why record the acquisition time, URL, and collection method?
- To prove every statement in the post is true
- To ensure that a page cannot later change
- To show what content was available and how it entered the investigative record
- To replace legal review of the search
- A sentiment model returns “high anger” for a political post. What is the safest interpretation?
- Proof that the author plans violence
- A validated measure of the author’s mental state
- Authority to add the person to a watch list
- A model estimate that should not be treated as intent or a threat without review
- Which evaluation better tests social-media monitoring quality?
- Measure identity errors, missed relevant posts, translation mistakes, and irrelevant alerts separately
- Count total posts collected
- Compare the analyst’s workload before and after launch only
- Use vendor-selected examples alone
- How should an AI-generated thread summary be used?
- As a substitute for preserving source content
- As a navigation aid checked against the original posts and replies
- As a verified witness statement
- As a reason to retain all unrelated personal information