In advance of the World Passenger Festival 2023, I was asked to answer a few questions. These were published in a document for those attending. (Question 1 is here , Question 2 is here, Question 3 is here and Question 4 is here)
Here's the fifth question.
AI is the technology of the moment: do you think the
public transport sector is taking the most of the opportunity with this
technology? What do the possibilities look like?
It is pretty hard to avoid the mass of AI related articles
and opinions right now. With Artificial Intelligence ‘experts’ seemingly
appearing from every direction (especially on LinkedIn!).
But it is obvious that our sector is not currently making
full use of this technology. Sure, some more innovative companies have been
using Machine Learning (ML) processes for some while, such as to ‘watch’ video
clips & automatically report safety issues or to cleverly merge disparate
transport data sources to analyse & enrich them. But transport is already
late to the AI party in many ways.
As an example, the current generation of Generative AI services (e.g. Chat GPT & Google Bard) learn by taking in huge amounts of content and processing it to create in-depth replies to a range of questions. But these services are limited at scraping documents & websites, so the best way to train them is to provide them with access to APIs of data sources. But in a sector where the zipped-up text file is often the most popular data exchange format … we are missing the technological opportunity to train these services properly and then learn how to use them for better public transport.
No comments:
Post a Comment