IB Computer Science HL · Year 2 · Lesson 55

Ethics retrieval + 2027 case study synthesis + final Paper 1 consolidation

Machine-learning decisions have ethical consequences involving bias, privacy, transparency, accountability and social impact. Machine learning systems learn patterns from data; the learning approach, application and hardware shape what the system can do.

Paper 150 minutesA4.4.1 · A4.4.2 · A4.1.1 · A4.1.2 · A4.2.1 · A4.2.2 · A4.2.3…
Use Next or the arrow keys.
Course homeIA hub2027 case study
Today’s targets

What you need to be able to do

2 / 8
  • DiscussThe ethical implications of machine learning in real-world scenarios
  • DiscussEthical aspects of the increasing integration of computer technologies into daily life
  • RetrieveReconnect today’s new material to previously taught content from this topic.
A4.4.1A4.4.2A4.1.1A4.1.2A4.2.1A4.2.2A4.2.3A4.3.1A4.3.2A4.3.3A4.3.4A4.3.5A4.3.6A4.3.7A4.3.8A4.3.9A4.3.10
Paper 1 lensMatch the depth of every response to the command term. Previously learned content can move quickly, but retrieval must still be accurate.
Page 2 of 8
Retrieve

Rapid Recall Deck

3 / 8

Say the answer aloud before flipping. Mark secure knowledge quickly and spend time on the gaps.

  • Which ethical issues should you consider when evaluating machine learning?: Accountability, algorithmic fairness, bias, consent, environmental impact, privacy, security, societal impact, and transparency.
  • Why can bias in training data create ethical problems?: A model can learn and reproduce skewed or exclusionary patterns, leading to systematically inaccurate or unfair outputs for some groups or situations.
  • Why must ethical guidelines be reassessed as technology changes?: New capabilities and uses create new risks and trade-offs that older rules may not anticipate, including effects on rights, privacy, and equity.
  • What is conditional image generation?: Image generation guided by a specified input such as a class label, sketch, segmentation map, text prompt, or source image so the output follows given conditions.
  • What is unconditional image generation?: Image generation without a predefined conditioning input, relying on patterns learned from the training data to create new outputs.
  • How does a diffusion model generate an image?: It begins from noise and uses an iterative denoising process—typically with a neural denoiser such as a CNN—to progressively reconstruct a coherent image.
  • What is a DDPM?: A denoising diffusion probabilistic model that formalizes the gradual addition and removal of noise used in diffusion-model training and generation.
  • What are the generator and discriminator in a GAN?: The generator turns a D-dimensional noise vector into synthetic images; the discriminator compares real and generated images and predicts whether they are real or fake.
  • What is mode collapse in a GAN?: A training failure where the generator produces only limited varieties of output instead of representing the diversity of the training data.
  • What is latent space in a VAE?: A compressed, organized representation where similar images are located near one another, allowing generation or modification by moving through that space.
  • What is distinctive about a flow-based model?: It learns an exact reversible mapping between random noise and realistic images, supporting generation, reproducibility, and traceability of how an image was formed.
  • Which factors should Visionary Studios use to compare generative-AI models?: Output quality, computational efficiency, training stability, flexibility/scalability, consistency, plus ethical/legal concerns such as dataset curation, intellectual property, bias/fairness, and AI disclosure.
  • Which five machine-learning approaches should be secure by the final lesson?: Supervised learning, unsupervised learning, reinforcement learning, transfer learning, and deep learning.
  • What hardware options can support demanding machine-learning workloads?: GPUs, TPUs, ASICs, FPGAs, edge devices, cloud platforms, and HPC centres, selected according to processing, storage, scalability, latency, and power needs.
Page 3 of 8
A4.4.1 + A4.4.2 · Learn

Core knowledge and application

4 / 8
A4.4.1Discuss

The ethical implications of machine learning in real-world scenarios

Machine-learning decisions have ethical consequences involving bias, privacy, transparency, accountability and social impact.

Exam moveGive a balanced, evidence-based treatment of different sides before concluding.
Required detail 1Ethical issues may include accountability, algorithmic fairness, bias, consent, environmental impact, privacy, security, societal impact, transparency
Required detail 2The challenges posed by biases in training data
Required detail 3The ethics of using machine learning in online communication may include concerns about misinformation, bias, online harassment, anonymity, privacy

Explain it without notes

Discuss: The ethical implications of machine learning in real-world scenarios in the context of an organization deploying an ML system that affects real people.

  • Ethical issues may include accountability, algorithmic fairness, bias, consent, environmental impact, privacy, security, societal impact, transparency
  • The challenges posed by biases in training data
  • The ethics of using machine learning in online communication may include concerns about misinformation, bias, online harassment, anonymity, privacy
A4.4.2Discuss

Ethical aspects of the increasing integration of computer technologies into daily life

Machine-learning decisions have ethical consequences involving bias, privacy, transparency, accountability and social impact.

Exam moveGive a balanced, evidence-based treatment of different sides before concluding.
Required detail 1The importance of continually reassessing ethical guidelines as technology advances
Required detail 2The potential implications of emerging technologies such as quantum computing, augmented reality, virtual reality and the pervasive use of AI on society, individual rights, privacy and equity

Explain it without notes

Discuss: Ethical aspects of the increasing integration of computer technologies into daily life in the context of an organization deploying an ML system that affects real people.

  • The importance of continually reassessing ethical guidelines as technology advances
  • The potential implications of emerging technologies such as quantum computing, augmented reality, virtual reality and the pervasive use of AI on society, individual rights, privacy and equity
Page 4 of 8
Apply

Transfer to a new scenario

5 / 8
ScenarioAn organization deploying an ml system that affects real people needs a design or technical decision related to today’s topic. Explain what matters and why.
  • Ethical issues may include accountability, algorithmic fairness, bias, consent, environmental impact, privacy, security, societal impact, transparency
  • The challenges posed by biases in training data
  • The importance of continually reassessing ethical guidelines as technology advances
  • The potential implications of emerging technologies such as quantum computing, augmented reality, virtual reality and the pervasive use of AI on society, individual rights, privacy and equity
Page 5 of 8
Exam lens

Paper 1 practice

6 / 8
Build the response before checking notesUse precise terminology and match the required depth.
  1. Discuss: The ethical implications of machine learning in real-world scenarios in the context of an organization deploying an ML system that affects real people.
  2. Discuss: Ethical aspects of the increasing integration of computer technologies into daily life in the context of an organization deploying an ML system that affects real people.
Self-checkAnswer the exact command term. For explain, include mechanism/reason; for compare, pair criteria; for discuss/evaluate/justify, build supported reasoning and a conclusion.
Page 6 of 8
2027 Case Study

Generative AI research checkpoint

7 / 8

Build a reusable evidence bank for Paper 1 Section B. Keep claims technical, specific and supported.

Research checkpoint

Synthesize all four official HL challenges into a comparative evidence bank for Paper 1 Section B.

Challenge focus: Evaluate how a hybrid design could combine strengths and offset weaknesses of diffusion models, GANs, VAEs and flow-based models.

Denoising diffusion probabilistic model (DDPM)A diffusion framework that learns a gradual reverse process from noise toward structured data.
Know and apply in case-study context
Image-to-image translationTransforming one image representation into another, such as sketch-to-rendered image.
Know and apply in case-study context
D-dimensional noise vectorA vector of random values with D dimensions that can serve as input to a generative model such as a GAN.
Deep for HL where tied to GAN/hybrid challenges
Variational autoencoder (VAE)A generative model that learns a structured latent representation from which new samples can be generated.
Broad understanding sufficient

Four HL challenge areas

  1. Explain how iterative denoising creates images and evaluate the computational demands of the process.
  2. Evaluate training-data choices with particular attention to intellectual property, bias and mitigation.
  3. Explain why GAN training depends on a workable balance between generator and discriminator performance.
  4. Evaluate how a hybrid design could combine strengths and offset weaknesses of diffusion models, GANs, VAEs and flow-based models.
Paper 1 Section BPrepare short AO1/AO2 knowledge/application responses and a balanced AO3 evaluation supported by research.
Page 7 of 8
Homework

Finish the learning cycle

8 / 8

Exam preparation — main task

Complete targeted 2027 case-study research and cumulative Paper 1/Paper 2 practice. Record evidence and technical vocabulary you can use in extended responses.

Retrieval

Repeat today’s recall deck and revisit any item marked ‘Review again’.

Designed by David Xu
© 2026 David Xu. All rights reserved.
Educators and students are welcome to use these materials for non-commercial teaching and learning with attribution. Please share the original link when possible. Reposting, redistributing modified copies, removing attribution, or commercial use requires prior permission.
End of Lesson 55.