Trustworthy AI: Accuracy, explainability and interpretability, privacy, reliability, robustness, safety, and security or resilience to attacks


To use AI in human-critical applications that could affect well-being, environment, lifestyle, etc., we need to trust the AI systems and those who develop and implement them. The National AI Initiative notes:

“To be trustworthy, AI technologies must appropriately reflect characteristics such as accuracy, explainability and interpretability, privacy, reliability, robustness, safety, and security or resilience to attacks – and ensure that bias is mitigated.”

This vertical will focus on these characteristics, highlighting their technical and business challenges. The following definitions have been adapted from many sources [1] [2] [3].

– Explainability and Interpretability. Understanding is a requirement for developing trust. Explanations are necessary to enhance understanding, trust, and informed decision making. Explainability is the extent to which the internal mechanics of a model can be explained in human terms. Interpretability is about the extent to which a cause and effect can be observed within a system.

– Transparency. The AI data, system, and business models must be transparent. Humans must be aware when they are interacting with an AI system, and must be informed of its capabilities and limitations. Ultimately, transparency improves traceability and accountability.

– Privacy, Governance Risk, and Compliance. Besides providing privacy and data protection, data governance mechanisms must also be used, to guarantee data quality and integrity, while ensuring legitimized access to data.

– RobustnessSecurity/Resilience to attacks. ML algorithms could be vulnerable to novel adversarial inputs. Trustworthy AI systems must have Robustness (the ability to withstand the effects of adversaries), including adversarial threats and attacks.

– Fairness/Bias. Hidden biases could lead to discrimination and exclusion of underrepresented/vulnerable groups. To prevent this, AI systems must include proper safeguards against bias and discrimination. Except for when the bias is intrinsic to the ML algorithm (Algorithm Bias), biases are usually introduced in the ML model through its training data set: Sample BiasPrejudice BiasMeasurement Bias, and Exclusion Bias.



  • Bridging the principles to practice gap in data and AI ethics for business who wish to build trustworthy technology with positive social impact (AI for Good)
  • Building public trust in technology through public engagement and co-ideation with businesses and academics – societal informed research.
  • Accountability and Responsibility of tech.
  • Data privacy (GDPR), data sharing in the public sector, the role of synthetic data, consent on how and why data issued, the changing face of personal data. What is trusted data sharing? (Privacy preserving data mining)
  • Legislation – dynamic – being ahead of the curve
  • Environmental impact v sustainability (red and green AI)


  • Building public trust through explainable AI to the public – beyond algorithms and models
  • Upskilling SMEs on doing AI ethical (bias, fairness, data representations, algorithmic transparency, interpretability etc.)
  • Red and Green AI (Environmentally friendly and sustainable AI)
  • Explainability V Transparency V Intellectual Property Rights
  • What does a Code of AI Ethics and Governance look like?


  • Lack of skills, resources (time and money), lack of access to real world data
  • Ethical and responsible approach not embedded into business plan (cost effectiveness – improves reputation).
  • Accountability dimensions: moral, social, legal, ethical, political, economic, environmental (multi-objective and fuzzy)


  • P7000 standards – accessibility
  • The IEEE Ethics certification program.
  • How to integrate ethics into the technical data science / ML pipeline.
  • Data privacy (state of the art solutions especially with personal data), data sharing, the role of synthetic data.
  • Current dos and don’ts (i.e, GDPR) with regards to the use of data in AI model building.
  • How to build public trust through co-ideation and co-production (i.e., citizen juries)
  • The role of diversity and inclusion in tech ethics – making solutions more inclusive.


Keeley Crockett (Manchester Metropolitan University)
Catherine Huang (Google)
Rajesh Murthy (GAPASK Inc.)
Andreas Nuernberger (Otto-von-Guericke-Universität Magdeburg)
Francesco Flammini (Mälardalen University)
Marcello Ienca (EPFL CDH-DIR)
Alessandro Facchini (AI Labs IDSIA)
Christian Wagner (Nottingham University)