scieee Science in your language
[en] (orig)

23 TrainingThings for AI-assisted instructional design

Author: Manske, Antje; Feichtinger, Michael
Publisher: Zenodo
DOI: 10.5281/zenodo.17237958
Source: https://zenodo.org/records/17237958/files/23TrainingThings_AI_13102025.pdf
23 T ainingThings
AI-assis ed ins uc ional design
1. B ains o m cou se ideas e.g. by c ea ing p o-
and con lis s o di e en opics.
2. Resea ch abou a opic / a ge g oup and
ind esou ces e.g. by using pe plexi y.ai o
Googles No ebook LM.
3. Iden i y po en ial cons ain s & challenges.
9. De elop ealis ic scena ios, examples and
p oblems a ge ed o you audience.
10. Check w i en con en o ypog aphy,
g amma & conciseness.
11. Ge eedback (e.g. Did I miss some hing? A e
he concep s in oduced in logical o de ?).
12. Analyse he i o cou se ma e ial o a goal
and a ge g oup.
AI will no eplace you as an ins uc ional designe s —bu using La ge Language Models can help
you make a cou se mo e ele an , speed up design, and enhance lea ning wi h ac i e and social
me hods. Read page wo o some ecipes o c ea e you own p omp s.
This guide p esen s ou own bes p ac ices and does no assume o co e all possible aspec s.
Always e alua e AI ou pu s - and apply you own skills and c ea i i y!
13. B eak down p oblems in o manageable
sub- asks wi h inc easing complexi y.
14. C ea e mul iple-choice ques ions & good
dis inguishe s (e.g. C ea e 8 answe
op ions & ma k he co ec one wi h a *).
15. Enhance accessibili y by c ea ing image
desc ip ions o ansc ip s.
16. Design gami ied ac i i ies, scena io-based
and ole-playing ac i i ies ailo ed o
cou se objec i es.
17. C ea e images (e.g. DALL-E / Midjou ney)
o desc ibe exis ing images in aining
ma e ial and c ea e simila ones.
18. C ea e ideo sc ip s.
19. C ea e cou se announcemen s and
pa icipan communica ion.
20. C ea e a ge ed “Wha is in i o me”
messages o he cou se based on he
cou se ma e ial.
21. P oduce epo s based on lea ning analy ics
and eedback shee s.
22. Ca ego ize ex answe s in eedback o ms.
23. C ea e ecommenda ions o he nex ound
o cou se upda es.
E alua ion & Repo ing
4. Gene a e lea ning objec i es on di e en
le els o Blooms Taxonomy.
5. Use AI o in o m decisions, e.g. abou
app op ia e cou se o ma s.
6. Gene a e ou lines & agendas.
7. B ains o m ac i i y ideas o iceb eake s,
social lea ning, mo i a ing pa icipan s, …
8. Choose he igh o ma o quizzes,
assessmen s and p og ess checks based on
lea ning objec i es.
Implemen a ion
DOI: 10.5281/zenodo.17237958
Needs Analysis
Cou se Design
Cou se De elopmen
Media & Ac i i y C ea ion
23 T ainingThings
AI-assis ed ins uc ional design
When c ea ing you own p omp s - he e a e some simple p omp amewo ks you can use o
op imize hem:
In addi ion, you can use ooms / p ojec s o „cus om GPTs“ o make he esul s mo e accu a e,
consis en and o no ha ing o epea ins uc ions:
•e.g. “gi e me pe cen age con idence sco es o all ou pu s and c ea e ables wi h p os and
cons when p o iding di e en al e na i es”
In hese specialized cha s, you can upload cou se ma e ial and o he documen s o p o ide he model
wi h mo e knowledge. In mos ools, you can choose in you se ings i you inpu will be used o u he
ain he public model. Be ca e ul o emo e any pe sonal da a.
Las bu no leas , Me a P omp s can gi e you AI-gene a ed ad ice how o imp o e you p omp s.
He e a e some examples:
•“Sugges 3 be e ways o ph ase his p omp using he STAR me hod”
•“My goal is o … How can you help me sol e his ask?”
S - SITUATION (con ex / p oblem o
backg ound in o ma ion)
T - TASK (objec i e / goal o be achie ed
wi h he p omp )
A - ACTION (conc e e s eps / how)
R - RESULT (speci ic
ou pu equi emen s)
C -CONTEXT (wha you a e ying o do &
o whom)
A -ASSIGNMENT ( he ask o he LLM)
F -FORMAT ( ex , able, # o
al e na i es,..)
E -EVALUATE (check sou ces,
ou pu , biases,..)
Full Example P omp s:
C ea e Lea ning Objec i es
“I wan o c ea e a cou se abou da a managemen planning o PhD s uden s (no p e ious knowledge).
The goal is o aise awa eness o he bene i s o DMPs o hem and enable hem o c ea e hei DMP wi h
suppo om a Da a S ewa d. C ea e 7 cou se objec i es on di e en le els o Blooms Taxonomy.”
Ge eedback
“Think like an ins uc ional designe and gi e me eedback on he cou se ma e ial uploaded wi h ega ds
o lea ne ac i i y, exe cises ..”
B ains o m ac i i y ideas
“Fo my DMP cou se o PhD candida es, I need some ac i i y ideas o he in oduc o y chap e s and
lowe -le el lea ning goals (wi h ega ds o Blooms Taxonomy). These objec i es a e…”
C ea e scena ios and examples
“Come up wi h i e ic ional scena ios o esea che eques s o he deposi ion o hei esea ch da a ha
a e di ec ed a esea ch suppo s a . They should be om di e en disciplines.”