BCS has grouped the ‘ideas for good’ into three main strands: Utopian visions, where we are looking for folks to push the boat out and dream; Practical suggestions, looking at what annoys you right now, and asking you to suggest what could we do about it; and Re-imagining - a starting-from-scratch type approach, asking what we take for granted that we could rip up and start again.
Practical suggestions: Get your AI here
John Peto suggests a new AI, optimised to explain other AI neural network decisions.
‘As (weak) AI,’ he says, ‘becomes ubiquitous, it’s going to make lots of mistakes and we’re going to need to understand why it makes mistakes so that we can further optimise. Personal AI is needed to combat corporate AI so that we’re not constantly taken for a ride.’
He uses some examples:
- Anti-fraud AI. It’s obvious that I don’t want to make an out-of-character large payment to a newly opened bank account in a far-off place without some additional checks, like talking to a person. Also, I could do with some AI to help with phishing and click bait and other annoyances that are easily spotted.
- AI to manage vehicles. Equipped vehicles share routes evenly around traffic jams, so that winners and losers are rotated evenly. Today I might get sent down a lengthy detour, but tomorrow I’ll get paid back with a quick route that others have been diverted away from.
- AI to stop trucks hitting bridges and getting stuck down unsuitable roads: all the common-sense stuff that we still struggle with.
- AI to tell me if I could take the train cheaper, and maybe some constantly varying train tickets prices, to ensure that trains are used optimally, based on demand.
- Quantum computing regulation: so that the first owners don’t get a ridiculously unfair advantage in trading markets by uncovering trends and techniques which will most likely be followed by a crash. Just regulation to use new technology responsibly, not just for pointless money grabbing.
‘There’s a short wish list,’ he writes, ‘but I’d swap it all for one final wish: use AI and psychology and anything else at our disposal to make rampant and unsustainable consumerism as uncool as it undoubtedly is. We need to shame people into leading greener lives, myself included.
Sheep mentality means that we should be able to reverse a lot of the current trends with the same marketing psychology that we’ve used to create the problems. We just need to speed things up in this space.’
Re-imagining: education
Paul Chau sends us some ideas on assessment in school. ‘The standardised learning path has been working for a long time’, he says, ‘however, with the evolution of data science and usage of elearning, it may be possible to renew the whole education system to make children adaptive with the skills they need in the current information age.
‘This is how it works. First, if we do most or all teaching with an elearning system, the elearning system can collect the student behaviour throughout the teaching process, and the materials can then be adjusted by analysing the student behaviour on each particular topic and use the process itself to be an assessment instead of a single end-of-term examination.
‘Second, as the assessment method is renewed, teachers are able to provide teaching materials, not according to syllabus, but generate more chances for student-initiate learning. By analysing student participation status throughout the years, the system can then match the student with some career paths fitting the student ability and interest and thus minimise the time a student spends on unnecessary subjects/skills. That could help a student to find their own interests and useful suggestions on career planning, and also help the country to keep track of the quantity and quality of talents in each sector.
Have a view - get involved
