Taught sessions:
- Technical aspects of AI – What are artificial intelligence, machine learning, neural networks and deep learning? How do they relate to one another? An introduction to the three branches of machine learning: supervised, unsupervised and reinforcement learning, and their uses.
- Algorithmic Bias – How can machine learning algorithms reinforce and exacerbate existing biases? Are AI systems biased by design? How does the data we draw upon in training a system shape its outputs, and how can we anticipate and address algorithmic biases?
- Can we build feminist AI? – What do we mean when we say a chatbot is "feminist"? Can AI really ever be gender neutral? How can we use feminist theories and methods to develop AI technologies that respond to algorithmic bias, tackle existing social inequalities, and address power imbalances?
- Self driving cars and algorithmic decision making – How close are we to self-driving car technology? Will driverless vehicles fundamentally
change the transportation system and reshape our built environment? Can an algorithmic system make ethically sound decisions?
- AI in Healthcare – AI has been touted as a critical tool in responding to health crises such as the Covid-19 pandemic. But can we apply machine learning while preserving privacy? Has the pandemic facilitated an expansion of surveillance systems? How can health data become de-anonymised?
- Controlling the future of AI – The Control Problem is the challenge of aligning the goals of future superintelligent AI systems with human values such that future systems will help and not harm humanity. Why is AI safety such a vexatious problem? Why do AI systems produce perverse behaviours and undesirable outcomes, and what can we do to realign them?
Group activities and discussions:
In addition to the lectures above, this track will use interactive websites, group discussions, games and a film screening of Coded Bias to explore and explain how AI and machine learning works and the challenges to overcome to make AI less biased, more fair and protect individual privacy. We’ll explore:
- Using AI in interviews and hiring decisions
- No code introductions to machine learning and training machine learning models with your own content
- How to judge the moral dilemmas faced by machine intelligence – what would you do?
- What does Google know about you – and what can you do about it?
Independent small group learning and practice:
Working in groups of two, you will undertake FutureLearn and UAL Creative Computing Institute’s Design a Feminist Chatbot online course. The course involves further exploration of bias and feminist design principles, developing user research, storyboarding and prototyping skills as you develop a feminist chatbot prototype. The prototype is then applied to an existing chatbot code template. By the end of the week, you will have produced a small portfolio of work, including storyboards, user research artifacts and a working prototype.
Final project:
Learners will share their chatbot designs and discuss the challenges and how they addressed them with the wider group.
Participation is restricted to MISDI students so please make sure you are logged in to Training and Development System to make your booking.