AI is getting more accessible (as in more easily available or pushed-down-our-throat, depending on your perspective). My home Windows 11 laptop now has the ‘free’ version of copilot on it, in the taskbar, pretty obviously it’s also available in the Bing search and I’m expecting Apple Intelligence to drop onto my phone imminently (I believe it is imminent). In the last week I’ve noticed Google Gemini appearing on adverts all over the TV I watch. AI’s seem to be everywhere.
I’ve been watching the ‘AI revolution’ carefully since it exploded in Nov 2022. I’ve blogged about it before – both about the results of experimenting with it and my fears about it. My mood about it changes from day-to-day. I think there are some places it can make a real difference to our learning/teaching/assessment and media platforms, being really useful to staff and students and other places where it’s just a ‘shiny’ thing and is a bit of a gimmick and I wonder how long before users get bored and move to other things. I’ve written about the risks around use and what happens to your data and I’m more and more aware of the black-box nature of AI and how sometimes our suppliers don’t necessarily understand what happens inside that box.
One thing I’ve really been thinking about recently is about how much it’s now being pushed for ‘free’. Now, if I know one thing, I know things are rarely free. AI certainly can’t be free. It may look ‘free’ for now but it won’t be forever and the worry I have is about whether we will get the opportunity to choose whether or not to pay for it or not, I suspect we won’t. AI is massively expensive – huge data centres/processing resources like electricity and water. Suppliers can’t just absorb the massive costs – they may do so now but at some point they will need to recoup the money they’ve spent on the AI we’ve had access to and been using for ‘free’. So our prices will go up – even if we don’t use the AI tools provided because we don’t find them useful (ie they are a gimmick). If it looks like it will be ‘free’ forever, the supplier will likely ….be benefiting in some other way … like collecting and processing your data
We’ve done a lot of work in LTW on our SADIE (Scoping AI Developments in EdTech at Edinburgh) project. This is a small project we’ve used to look at some of the AI being offered and highlighting risks, tool value and environmental impact. We’ll be publishing more about this project over the next while. If you are a member or staff or a student at the University, you can read more about the project on our SADIE site.
I feel strongly we need more transparency around AI (and some suppliers are better than others, you know who you are!!) but also about the impact/cost which may be passed on at a later date. The Google Gemini adverts show someone having a jolly lovely chat with the AI about what it can do. There’s no small print to say what happens to the data you give it/how the quality of the response it gives might be nonsense/the fact that it may use 4 or 5 times the environmental impact to use an AI than to answer a question using a standard web search.
