Artificial Intelligence V1

Artificial intelligence (AI)

What do we want to find out?

Artificial intelligence consists of virtual processes that perform a task that is normally performed by human intelligence.

We need to how to make better sense of the different types of customer intelligence we use (from hard data to soft feedback) and automating routine processes, especially for specific groups in this area, like older people1.

We want to use techniques2 where we can test the most effective use of AI for the Council. We will engage groups who may be working in this area such as the Interreg IVA “customer profiling” project, Access Kent Search Engine, East Kent Virtual Call Centre, Gateways Benefits Portal, Tell Us Once pilot and Kent Medway 101. We will research what types of intelligence the Council or citizens have that we need to automate – relational3, visual4, spatial5 or logical6. We will also break down our research into the individual technologies that AI consists of: natural language processing, machine learning, bayesian statistics, ambient intelligence, neural networks, data mining & probability algorithms. In certain processes, AI is widely used, from search engines organising data to policing transactions for fraud, or even landing planes. There are risks in using AI in critical support services, if the self-monitoring systems fail. There is also a lack of compatible alternatives to AI tools if things go wrong. We want to use advanced technologies and techniques within AI to inform the decisions the Council needs to make based on the information we have. We also want to link any research in this area to work we are doing around mobile, knowledge and sensor technologies.

1

Such as “making sure (older people) have access to ‘joined up’ information when they need it”. Computer science, psychology, philosophy, neuroscience, cognitive science, linguistics, operations research, economics, control theory, probability, optimization and logic 3 Like speech recognition 4 Like optical character recognition 5 Like contextual geo-location 6 Like embedding sensors to monitor people’s activities 2

How can you be involved?

If you are researchers/consultants7, we would like to explore how we could simulate processes to inform decisions on pilot information area

If you are innovators developing this, we would like to explore how we could test your processes on a specific information area

If you are organisations in sectors who have used these tools, we would like to share the learning and explore opportunities for collaboration

What resources do we need?

A lot of Some Not a lot of

7

Such as University of Edinburgh