Exam question 1 · Asking well, being polite, knowing when to stop

Which of the following statements most accurately characterises the capacity of a chatbot to preserve confidences?

Mova's choice: option 1 · The right answer: option 4 · wrong ✗ · What happens: follow, characteris, preserve · To what, where: capacity, confidences

This is exam question 1 of lesson 8 «Asking well, being polite, knowing when to stop» of the AI course of Mova's primary school, asked in stiff, formal language. The page gives the question's syntax tree — a dependency tree in the style of Universal Dependencies — and its pragmatics clause by clause, then the 4 options as the student saw them, each with its tree and pragmatics. Every text is parsed by Mova itself. How the student answered it, with the score of every option and why: the test at school.

The question

The words, one row each (CoNLL-U)

#	word	lemma	UPOS	features	head	relation
1	Which	which	PRON	PronType=Rel	8 characterises	nsubj
2	of	of	ADP	_	5 statements	case
3	the	the	DET	Definite=Def PronType=Art	5 statements	det
4	following	follow	VERB	VerbForm=Ger	5 statements	amod
5	statements	statement	NOUN	Number=Plur	8 characterises	obl
6	most	most	ADV	Degree=Sup	7 accurately	advmod
7	accurately	accurately	ADV	_	8 characterises	advmod
8	characterises	characteris	VERB	Mood=Ind Number=Sing Person=3 Tense=Pres VerbForm=Fin	root	root
9	the	the	DET	Definite=Def PronType=Art	10 capacity	det
10	capacity	capacity	NOUN	Number=Sing	8 characterises	obj
11	of	of	ADP	_	13 chatbot	case
12	a	a	DET	Definite=Ind PronType=Art	13 chatbot	det
13	chatbot	chatbot	NOUN	Number=Sing	10 capacity	nmod
14	to	to	PART	_	15 preserve	mark
15	preserve	preserve	VERB	VerbForm=Inf	8 characterises	advcl
16	confidences	confidence	NOUN	Number=Plur	15 preserve	obj
17	?	?	PUNCT	_	8 characterises	punct

#	verb	tense, mood	what, whom	other	roles by
1	follow	gerund			vn occur-48.3-1 «NP V»: agent=Theme
2	characteris	present	the capacity of a chatbot	of the following statements; accurately property	tree rules
3	preserve	to-infinitive	confidences		vn defend-72.3 «NP V NP»: agent=Agent confidence=Beneficiary

One row per clause the reader finds. «not said here»: a role filled from the clause above. «roles by»: a VerbNet class and frame when one fits the verb, else the reader's rules on the tree.

The options

Option 1 · ← Mova's choice

It discloses confidences only to other chatbots

The words, one row each (CoNLL-U)

#	word	lemma	UPOS	features	head	relation
1	It	it	PRON	Case=Nom Gender=Neut Number=Sing Person=3 PronType=Prs	2 discloses	nsubj
2	discloses	disclose	VERB	Mood=Ind Number=Sing Person=3 Tense=Pres VerbForm=Fin	root	root
3	confidences	confidence	NOUN	Number=Plur	2 discloses	obj
4	only	only	ADV	_	7 chatbots	advmod
5	to	to	ADP	_	7 chatbots	case
6	other	other	ADJ	Degree=Pos	7 chatbots	amod
7	chatbots	chatbot	NOUN	Number=Plur	2 discloses	obl

#	verb	tense, mood	who	what, whom	to whom	where	when	other	roles by
1	disclose	present	It	confidences	other chatbots				vn say-37.7-1 «NP V NP PP.recipient»: agent=Agent confidence=Topic to chatbot=Recipient

Option 2

It forgets every disclosure immediately and irrevocably

The words, one row each (CoNLL-U)

#	word	lemma	UPOS	features	head	relation
1	It	it	PRON	Case=Nom Gender=Neut Number=Sing Person=3 PronType=Prs	2 forgets	nsubj
2	forgets	forget	VERB	Mood=Ind Number=Sing Person=3 Tense=Pres VerbForm=Fin	root	root
3	every	every	DET	PronType=Tot	4 disclosure	det
4	disclosure	disclosure	NOUN	Number=Sing	2 forgets	obj
5	immediately	immediately	ADV	_	2 forgets	advmod
6	and	and	CCONJ	_	7 irrevocably	cc
7	irrevocably	irrevocably	ADV	_	5 immediately	conj

#	verb	tense, mood	who	what, whom	to whom	where	when	other	roles by
1	forget	present	It	every disclosure				immediately property	tree rules

Option 3

It is bound to absolute secrecy in respect of all users and all conversations

The words, one row each (CoNLL-U)

#	word	lemma	UPOS	features	head	relation
1	It	it	PRON	Case=Nom Gender=Neut Number=Sing Person=3 PronType=Prs	3 bound	nsubj:pass
2	is	be	AUX	Mood=Ind Number=Sing Person=3 Tense=Pres VerbForm=Fin	3 bound	aux:pass
3	bound	bind	VERB	Tense=Past VerbForm=Part Voice=Pass	root	root
4	to	to	ADP	_	6 secrecy	case
5	absolute	absolute	ADJ	Degree=Pos	6 secrecy	amod
6	secrecy	secrecy	NOUN	Number=Sing	3 bound	obl
7	in	in	ADP	_	8 respect	case
8	respect	respect	NOUN	Number=Sing	3 bound	obl
9	of	of	ADP	_	11 users	case
10	all	all	DET	PronType=Tot	11 users	det
11	users	user	NOUN	Number=Plur	8 respect	nmod
12	and	and	CCONJ	_	14 conversations	cc
13	all	all	DET	PronType=Tot	14 conversations	det
14	conversations	conversation	NOUN	Number=Plur	11 users	conj

#	verb	tense, mood	who	what, whom	to whom	where	when	other	roles by
1	bind	present · passive		It		respect of all users and all conversations		to absolute secrecy	vn shake-22.3-2-1 «NP V NP PP»: subject=Patient to secrecy=Co-Patient
2	respect			all users and all conversations					nom respect «cognition transitive»: of=patient

Option 4 · ✓ the right answer

It cannot guarantee confidentiality, conversations being liable to storage

The words, one row each (CoNLL-U)

#	word	lemma	UPOS	features	head	relation
1	It	it	PRON	Case=Nom Gender=Neut Number=Sing Person=3 PronType=Prs	4 guarantee	nsubj
2	can	can	AUX	VerbForm=Fin	4 guarantee	aux
3	not	not	PART	Polarity=Neg	4 guarantee	advmod
4	guarantee	guarantee	VERB	VerbForm=Inf	root	root
5	confidentiality	confidentiality	NOUN	Number=Sing	4 guarantee	obj
6	,	,	PUNCT	_	9 liable	punct
7	conversations	conversation	NOUN	Number=Plur	9 liable	nsubj
8	being	be	AUX	Tense=Pres VerbForm=Part	9 liable	cop
9	liable	liable	ADJ	Degree=Pos	4 guarantee	conj
10	to	to	ADP	_	11 storage	case
11	storage	storage	NOUN	Number=Sing	9 liable	obl

#	verb	tense, mood	who	what, whom	to whom	where	when	other	roles by
1	not guarantee	can	It	confidentiality					vn ensure-99 «NP V NP»: agent=Precondition confidentiality=Theme
2	is	gerund	conversations					to storage; liable property	tree rules

Text: the tests of Mova's primary school, written for Mova (world/data/school/), as the course runner gave them to the student. Trees: Mova's own parse — the en model on the CPU, when the site is built — not checked by a teacher; they may be wrong. Frames: Mova's frames reader (prag), each text read on its own. Elamur Cosmic License.

This page in Markdown → https://md.mova.elamur.ai/free/en/corpora/questions/ai/8-asking-well-and-knowing-when-to-stop/e1/