Text (Instruction Step Text)
From The Embassy of Good Science
Describe the actions the user should take to experience the material (including preparation and follow up if any). Write in an active way.
- ⧼SA Foundation Data Type⧽: Text
1
Taking part in research: A guide for citizen scientists on environmental and climate considerations +
'''1. Project purpose and who is behind it'''
You should receive a clear, short description of the project’s aims (for example, monitoring local heat islands or tracking biodiversity change), who coordinates the project (institutions), who funds it, and how the results are expected to be used (e.g. informing local policies, contributing to scientific publications, supporting environmental management). This information should be provided in accessible, non technical language. Where relevant, the project should also explain whether its activities or results could create environmental or social impacts beyond the immediate study context, including effects on other communities or future generations, and what measures are taken to minimise such risks
The European Green Deal is the European Union's comprehensive growth strategy launched in 2019, aiming to transform the EU into a modern, resource-efficient, and competitive economy. Its central goal is to make Europe the first climate-neutral continent by 2050, with an interim target to reduce net greenhouse gas emissions by at least 55% by 2030. The framework covers energy, industry, agriculture, transport, and biodiversity to create a sustainable, inclusive, and clean, green transition.
'''2. Your role as a citizen scientist'''
The project should explain clearly what you will do, such as making observations, uploading photos or sensor data, joining workshops, or helping interpret results, including activities that may involve natural areas or environmental data
You are free to decide what level of involvement is realistic for you, and you are not obliged to take on tasks beyond what you agreed at the start. You can also adjust your level of involvement over time or decide to stop at any time.
Do No Significant Harm" (DNSH) is a European Commission principle ensuring investments, particularly under the Recovery and Resilience Facility, do not cause significant environmental damage. It requires projects to meet criteria across six objectives, including climate change mitigation, adaptation, water protection, circular economy, pollution control, and biodiversity
'''3. What happens with the data you contribute'''
The project should explain which types of data will be collected including, where relevant, data that may be collected about you and data that you will collect or generate through your participation (for example dates, locations, measurements, photos, and any environmental or climate-related information you might gather). You should be informed on how to ask questions and feel comfortable asking them if anything is unclear. You should know where and for how long the data you will provide will be stored, how it will be protected and whether it will be openly available or shared under conditions. You should also know how you will receive feedback on results (e.g. maps, dashboards, newsletters or meetings). You should feel comfortable asking questions if anything is unclear.
'''4. Terms, conditions and duration'''
You should receive clear terms and conditions that explain how you can participate, how the data you contribute can be used, shared or reused, under which licence, and how long the project and its data infrastructure are expected to run. You should have an opportunity to read these and ask questions before you decide whether they are acceptable to you.
The practical identification of AI ethics issues requires applying the seven Assessment List for Trustworthy Artificial Intelligence (ALTAI) requirements to concrete operational settings using the AIOLIA Issues Identification Checklist. The evaluation process begins by Establishing Operational Context (Section 0), where practitioners define system parameters—including input/output specifications, user profiles, decision autonomy, and affected stakeholder groups—before assessing risk. Grounding risk identification in specific operational contexts prevents vague, generic evaluations and ensures that downstream risk assessments reflect actual deployment realities.
Across Sections 1–7 of the checklist, practitioners systematically evaluate real-world scenarios to detect failure modes and ethical vulnerabilities. Evaluated risks include subtle loss of student agency caused by over-reliance on adaptive AI tutoring prompts, severe safety hazards stemming from data poisoning in industrial monitoring tools, and privacy violations arising from unverified training data in organizational assistants. The evaluation further examines black-box opacity in automated risk scoring platforms, algorithmic bias and proxy discrimination in content moderation systems, skills degradation among domain experts, and unclear legal liability during clinical decision-support failures.
To complete the identification process, qualitative observations and Likert-scale evaluations are synthesized into Structured, Auditable Issue Statements. Each statement explicitly documents the affected ALTAI requirement, the precise operational context, the root cause of the vulnerability, and the potential severity of harm. This standardized documentation creates a clear audit trail, establishing the necessary empirical foundation for selecting targeted technical and organizational mitigation measures. +
An AI system, as defined in the EU AI Act, provides the foundational reference point for determining the scope of AI governance and related ethics obligations. This definition is crucial because it establishes clear criteria for distinguishing AI-based technologies from other software systems, ensuring consistent interpretation across research, development, and policy contexts. A precise understanding of this definition helps stakeholders assess whether a system is subject to specific legal, ethical, and compliance requirements under EU rules. According to the definition of an AI system in the EU AI Act:
''‘AI system’ means a <u>machine-based system</u> that is designed to operate with <u>varying levels of autonomy</u> and that may exhibit <u>adaptiveness</u> after deployment, and that, for <u>explicit or implicit objectives</u>, <u>infers</u>, from the input it receives, how to generate <u>outputs</u> such as predictions, content, recommendations, or decisions that can <u>influence physical or virtual environments</u>.''
Source: European Union. (2024). Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 on artificial intelligence (EU AI Act). Official Journal of the European Union. Article 3(1) <u>https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32024R1689</u> +
Artificial Intelligence (AI) systems process data to recognize patterns, make decisions, and execute actions. At the foundation lies Machine Learning, which enables systems to learn from data to identify patterns and make predictions without manual rule coding, operating through supervised, unsupervised, and reinforcement learning paradigms. Neural Networks and Deep Learning utilize brain-inspired layered architectures to process complex structures like images, speech, and text, though their multi-layered design introduces explainability challenges, often operating as a "black box". Generative AI and Large Language Models (LLMs) scale these capabilities to produce novel content across text, code, images, and media; however, they require critical oversight due to risks such as non-factual "hallucinations" and embedded data bias. Finally, the evolution from task-specific AI Agents to autonomous Agentic AI introduces systems capable of proactive planning and goal-setting, making robust governance, accountability, and human oversight essential safeguards.
===='''Data - The Fuel That Powers AI'''====
Data serves as the foundational fuel for all Artificial Intelligence systems. The overall quality, quantity, representative diversity, and fairness of training datasets directly govern how accurately and equitably an AI model behaves. Safeguarding sensitive information, ensuring broad representation, and eliminating dataset biases are critical requirements to prevent automated systems from reinforcing unfair decisions.
===='''Machine Learning — How AI Learns from Data'''====
Machine Learning allows computers to recognize underlying patterns and formulate predictions from data without requiring human developers to code every rule explicitly. It encompasses Supervised Learning (learning from labeled examples), Unsupervised Learning (discovering hidden patterns or clusters without labels), and Reinforcement Learning (learning through trial and error via rewards). Because models learn directly from historical examples, any embedded biases in training data will inevitably be reflected and potentially amplified in output decisions.
===='''Neural Networks & Deep Learning'''====
Deep Learning is a specialized subset of Machine Learning driven by Neural Networks—models inspired by the human brain that process inputs through layers of interconnected artificial neurons. These layered architectures enable advanced applications such as image recognition, speech processing, and text generation. However, their internal decision-making processes are highly complex, giving rise to the "black box" challenge where explaining the precise reasoning behind a specific output remains difficult.
===='''Large Language Models & Generative AI'''====
Generative AI utilizes patterns learned from vast datasets to generate entirely new content, including text, images, code, audio, and video. Large Language Models (LLMs) specifically leverage massive text corpora to synthesize fluent, human-like language. Despite their sophisticated outputs, these models lack true understanding and can produce "hallucinations"—outputs that appear highly convincing yet are factually incorrect or nonsensical—requiring continuous human verification.
===='''AI That Acts - From Agents to Agentic AI'''====
Artificial Intelligence is evolving from passive content generation to active execution, distinguished by varying degrees of autonomy. While traditional AI Agents operate reactively within a narrow scope to complete specific tasks under explicit instructions, Agentic AI exhibits proactive behavior by setting priorities, breaking down complex goals into multi-step plans, and taking independent initiative over time. As AI systems gain greater operational autonomy and require less step-by-step human guidance, the necessity for formal oversight, clear accountability, and ethical safeguards increases significantly.
When AI systems interact with human users, they directly influence cognition, decision-making, and behavior, shaping societal values, rights, and ethical norms. Ethics-by-Design is an approach that embeds ethical considerations into AI systems proactively from the very start of design and development, rather than attempting to rectify flaws post-deployment. Based on the European Commission's High-Level Expert Group guidelines for trustworthy AI, this methodology relies on seven core requirements: human agency and oversight, technical robustness and safety, privacy and data governance, transparency, diversity, non-discrimination and fairness, societal and environmental wellbeing, and accountability. Implementation spans three crucial layers—the Data Layer, Model Layer, and Deployment Layer. Failure to integrate ethical controls throughout these layers has led to severe real-world harms, including systemic algorithmic discrimination in hiring, risk assessment, healthcare allocation, and social benefit delivery. Consequently, Ethics-by-Design serves as a vital operational mechanism for complying with mandatory regulatory frameworks such as the EU AI Act (2024) and align with guidance like the ALTAI framework to ensure trustworthy AI. +
The European Union promotes Trustworthy Artificial Intelligence (AI) as the foundation for responsible AI development and deployment, requiring systems to be lawful, ethical, and technically robust. Established by the European Commission, the High-Level Expert Group on AI (HLEG) introduced the ''Ethics Guidelines for Trustworthy AI'' in 2019, defining four foundational ethical principles—Respect for Human Autonomy, Prevention of Harm, Fairness, and Explicability. These principles translate into seven practical requirements: Human Agency & Oversight, Technical Robustness & Safety, Privacy & Data Governance, Transparency, Diversity, Non-Discrimination & Fairness, Societal & Environmental Wellbeing, and Accountability. To operationalize these guidelines across the AI lifecycle, the EC created the Assessment List for Trustworthy AI (ALTAI). Moving beyond "box-ticking" checklists, the Ethics Readiness Levels (ERLs) framework utilizes semi-structured dialogues between technical and ethics experts to measure and advance ethical maturity across five discrete levels (ERL 0 to ERL 4). Applying these ethical frameworks is crucial when evaluating research projects against the regulatory context of the EU AI Act. While Article 2 provides a "safe harbor" research exemption for scientific R&D, real-world testing, pilots in sensitive operational settings (such as healthcare, education, or employment), and commercialization move projects into a "Grey Zone" where high-risk obligations or Article 5 prohibited practices may apply. +
What will I accomplish by taking this course? +
The AI Ethics Operationalisation Workflow bridges the gap between high-level ethics principles and concrete, day-to-day practices across the AI lifecycle. It is built on three core building blocks: '''Ethics Requirements''' (the 7 ALTAI requirements), '''Practical Measures''' (a catalogue of 122 technical and organizational safeguards), and '''Ethics Tensions''' ("right vs. right" conflicts between competing ethical priorities).
The framework executes through a '''5-Step Operational Workflow''':
#'''Establish Context:''' Define the system's role, data sources, users, and affected stakeholders using Section 0 of the Issues Identification Checklist.
#'''Identify Ethics Issues and Map them under ALTAI Requirements:''' Systematically evaluate risks across problem formulation, data engineering, model development, deployment, and monitoring using Sections 1–7 of the checklist.
#'''Propose & Prioritise Practical Measures:''' Browse, assess, adapt, and combine technical measures (for developers) and organizational measures (for management) from the Portfolio of Measures.
#'''Identify Ethics Tensions:''' Analyze and navigate unavoidable trade-offs between competing requirements, such as Privacy vs. Fairness or Technical Robustness vs. Transparency.
#'''Perform an Ethics Check:''' Validate findings and assemble auditable, traceable documentation of all decisions and residual risks.
This workflow applies directly to real-world scenarios, such as the '''PhiShield''' case study. When an automated system flags an engineer (Marcus) with an 89% phishing vulnerability score after failing a test during an emergency 14-hour shift, it demonstrates how context-blind evaluations lead to unfair promotion blocks. Resolving such incidents requires applying technical sensitivity analysis (e.g., MS85), establishing a formal Right to Challenge (e.g., MS92), and enforcing mandatory human oversight before taking adverse personnel actions. +
Close the exercise with underlining the importance of good communication in dealing with research integrity issues and dilemma’s. Continue with the next fragment or next part of the workshop. +
04 - Moral Case Deliberation: A Method for Analysing Cases in Research Ethics and Research Integrity +
Lastly, learning experiences and the outcome are evaluated. +
Ask the group to reflect on the process, and to evaluate if the learning objectives were met. Foster a brief dialogue on what might have been learned as a group.
In this step the facilitator may ask participants questions such as:
- Was it easy or difficult to identify the relevant principles and virtues in the chosen dilemma?
- Did this exercise help you with identifying and connecting to formally defined principles (e.g. from the European Code of Conduct for Research Integrity)?
- Did most of the players agree or disagree with the final choice?
- What were the main points of contention?
- Why did people disagree (e.g. differences in experience, training, background, values, norms…)?
- What were the other options?
- Was any alternative option proposed?
- Did anybody change her/his mind as a result of the discussion?
- Why would you NOT follow the morally ideal course of action?
- What is needed to act morally in your work setting? What were the most convincing arguments used in the discussion?
- On which areas do you feel there is insufficient consensus?
- How can you best address future dilemmas in your daily work?
- How can shared values and principles be fostered? +
This final part of the manual consists of two instructions, with the links listed below:
[https://public.3.basecamp.com/p/R5e8zxXRHwd27Mz5PPfooByh Certification]
[https://public.3.basecamp.com/p/vmLSq94iGyaNsbrWKFgFbCiN Recognition and networking] +
Katılımcıları sürecin geneli üzerine düşünmeye davet edin: bu oturumdan çıkardıkları dersler neler? Katılımcılara aşağıdaki soruları sorarak belirli sonuçlar çıkarmaya çalışın:
o Erdemler ve normlar arasında ilişki kurmak kolay mıydı yoksa zor muydu? Neden?
o Kendinizi vakayı sunan kişinin yerine koymanız erdemlere ve dolayısıyla norm ve davranışlara olan bakış açınızı genişletti mi?
Diğer katılımcıların belirlediği erdem ve normlar/ davranışlar sizin erdemlere daha farklı ya da geniş bir açıdan bakmanıza yardımcı oldu mu? Bunun uygulamada karşılaşacağınız AED ikilemleri karşısında düşünme şeklinizi etkileyeceğini düşünüyor musunuz? +
Gruptan genel olarak süreç üzerine fikir yürütmelerini ve bu alıştırma bağlamında öğrenme hedeflerinin karşılanıp karşılanmadığına ilişkin bir değerlendirme yapmalarını isteyin. Katılımcıları bu alıştırma ile neler öğrendikleri üzerine kısa bir diyalog yürütmeye yönlendirin.
Bu aşamada eğitmen katılımcılara aşağıdakilere benzer sorular sorabilir:
- Seçilen ikilem için ilgili prensip ve erdemleri belirlemek kolay oldu mu?
- Bu alıştırma sizin resmi olarak tanımlanmış prensipleri (ECoC) tespit edip bunlarla vakalar arasında bağlantı kurmanıza yardımcı oldu mu?
- Oyunu oynayan katılımcıların büyük çoğunluğu varılan nihai karara muvafakat etti mi?
- Anlaşmazlığa yol açan başlıca noktalar nelerdi?
- Katılımcıların bazı noktalarda hemfikir olmamasına sebep olan şeyler nelerdi (örn., kişilerin deneyimlerindeki, eğitimlerindeki, arka planlarındaki, değerlerindeki, normlarındaki vb. farklılıklar)
- Diğer seçenekler neydi?
- Herhangi bir alternatif seçenek önerildi mi?
- Tartışma sonucunda herhangi bir katılımcı fikrini değiştirdi mi?
- Ahlaki açıdan ideal olan şeyi YAPMAMANIZIN sebebi ne olurdu?
- Sizin iş ortamınızda ahlaki olarak iyi olana ulaşmak için neler gerekli?
- Tartışmada kullanılan en ikna edici argümanlar hangileriydi?
- Hangi noktalarda yeterince fikir birliğine varılmadığını düşünüyorsunuz?
- Gelecekte iş yaşamınızda bu gibi ikilemlerle en iyi hangi şekilde başa çıkabilirsiniz?
- Üzerinde daha yaygın bir şekilde anlaşmaya varılan değer ve ilkelere nasıl ulaşılır? +
Lade die Teilnehmenden abschließend ein, über den gesamten Prozess während der vergangenen Übung nachzudenken: Was ist für sie die Take-Home-Message, die sie aus dieser Übung mitnehmen? Versuche, einige Schlussfolgerungen oder Erkenntnisse festzuhalten, indem du die Teilnehmenden fragst:
- War es einfach, die Werte/Tugenden zu den Normen in Beziehung zu setzen? War es schwierig? Warum?
- Hat der Versuch, sich in die Lage der Person zu versetzen, die die Beispielsituation erlebt hat, deine Sichtweise auf Werte/Tugenden und damit auch auf Normen oder Verhaltensweisen erweitert?
- Haben die von anderen genannten Werte / Tugenden, Normen oder Verhaltensweisen dabei geholfen, anders über das Thema nachzudenken und zum Beispiel Werte / Tugenden anders oder umfassender zu betrachten? Wie wird diese Erfahrung aus der Übung heute dein Denken über Dilemmata im Forschungsalltag verändern? +
Invite participants to think about the entire process: what is the take home message of this session for them? Try to draw conclusions by asking participants:
o Was it easy or difficult to relate the virtues and norms to each other? Why?
o Did putting yourself in the case presenter’s shoes broaden the way you looked at virtues and, consequently norms and behaviors?
o Did the virtues and norms/behaviors identified by others help you to look at virtues differently or more broadly? Do you think that will influence your thinking on research integrity dilemmas in practice? +
[[File:Man overlooking view.png|center|frameless|600x600px]]
Kuhn suggested that all scientific knowledge is ‘situated’ knowledge and cannot represent a ‘view from nowhere’. We all view the world from within a particular set of social and epistemic practices. According to Kuhn, scientists working within different paradigms are effectively working in different worlds. But how do we know which paradigm we are working in? +
[[File:M10..png|center|frameless|600x600px]]
In terms of ethics dumping, the previously mentioned TRUST global code of conduct for equitable research partnerships offers a simple, jargon-free [https://www.globalcodeofconduct.org/the-code/ ethics code] comprised of 23 articles based around the moral values of Fairness, Respect, Care and Honesty, to help researchers ensure that international research is equitable and carried out without ‘ethics dumping’ or ‘helicopter research’.
In terms of AI ethics, we recommend consulting the Ethics of AI in Healthcare: A checklist for Research Ethics Committees which was developed by irecs colleagues, Alexei Grinbaum and Etienne Aucouturier at CEA (French Alternative Energies and Atomic Energy Commission), as well as the materials in the [https://classroom.eneri.eu/node/238 irecs AI and ethics module].
Chapter 5 of the [https://www.who.int/publications/i/item/9789240029200 World Health Organization’s Ethics and Governance of Artificial Intelligence for Health] outlines six key ethical principles for AI research in healthcare. These include protecting patient autonomy, promoting human wellbeing, ensuring transparency and explainability, fostering accountability, promoting inclusiveness and equity, and supporting AI that is both responsive and sustainable. These principles serve as essential reminders for researchers and policymakers to prioritise ethical considerations in the development and deployment of AI technologies in healthcare settings.
Another significant issue in the development of AI technologies across all fields is the potential for bias and inaccuracies in algorithms, which in the healthcare domain can result in incorrect diagnoses and treatment recommendations. These risks disproportionately affect vulnerable populations, raising concerns about inclusivity and equity. The [https://op.europa.eu/en/publication-detail/-/publication/d3988569-0434-11ea-8c1f-01aa75ed71a1 EU’s Ethics Guidelines for Trustworthy AI] emphasise that AI systems must be lawful, ethical, and robust throughout their Life cycle. This includes compliance with applicable laws, adherence to ethical principles, and ensuring technical and social robustness. Importantly, these guidelines call for mechanisms to prevent algorithmic bias and protect privacy. Unethical applications involving AI are defined as those that risk violating physical or mental integrity, create addiction, risk damaging social processes and public institutions (e.g. by social scoring or contributing to misinformation).
Projects must adhere to essential requirements, which encompass (but are not restricted to):
* People must be made aware that they are interacting with an AI system, its abilities and Limitations, risks and benefits.
* Mechanisms for human oversight, transparency and auditability must be built into the AI system.
* AI-systems must be designed to avoid bias in input data and algorithmic design.
* Compliance with data protection and privacy principles must be demonstrated.
Our hypothetical proposal is not seeking funding from Horizon Europe, however, the [https://www.bbmri-eric.eu/wp-content/uploads/The-Ethics-Appraisal-Scheme-_BBMRI-webinar-september-2021_version-for-dessimination.pdf EU ethics appraisal scheme (pp74-80)], provides relevant guidance for several concerns in this case study. It highlights the importance of transparency, requiring that individuals interacting with AI systems be fully informed about the system’s capabilities, Limitations, risks, and benefits. It also underscores the necessity of building human oversight, transparency, and auditability into AI systems, ensuring that AI development remains accountable and aligned with societal values.
Regulatory oversight has often lagged behind technological advancements, creating additional legal and ethical challenges. The WHO and EU guidelines, among others, stress the need for AI systems to comply with data protection and privacy principles, such as data minimisation, ensuring that only the necessary data is collected and used. This is crucial in building trust and safeguarding against the misuse of sensitive healthcare information.
It is important to remember that different guidelines and regulations will apply to research projects in order to comply with the requirements of different institutions, organisations and geographical locations. Listed in the further resources section are sources to explore on ethics dumping, some of the ethics committees in Africa and the current most relevant EU or international guidelines or standards related to AI in health and healthcare, but you may need to explore further afield to locate those that apply to different situations.
[[File:AI img8.png|center|frameless|600x600px]] +
[[File:Ge2Image8.png|center|frameless|600x600px]]
Treatments and therapies involving gene editing are already undergoing clinical trials for marketing approval in the EU and the US for certain diseases and are likely to incur equivalent costs to those of conventional gene-based therapies that are used for rare genetic diseases. However, they are very costly and may thus be restricted to wealthy patients or citizens in countries with corresponding health insurance or social security systems. The dilemma of resource allocation poses questions about the development of extremely expensive therapies. +
