AI and learning

Since the first public release of ChatGPT3 in November 2022, artificial intelligence (AI) has irrevocably changed the worlds of work and learning, along with huge impacts in almost every other sector imaginable.

Of course, it’s not just ChatGPT. There are several other providers, from large companies like Google and Amazon, as well as small, niche organisations all trying to carve out their section of this growing market.

I won’t delve into the mechanisms of how these AI engines work. That’s way beyond the scope of this article. But there are four key points that everyone needs to get to grips with:

1. AI is not intelligent. It’s “simply” statistics. Very complicated statistics, but just statistics. Based on massive amounts of existing data (see note below), it uses statistics to give the most likely output for the input you’ve given it.
2. AI uses vast amounts of energy. Each time you ask a question or request an image, the system trawls its treasure trove of existing data and builds something new for you.
3. AI is not infallible. The responses it gives are based on the data it has. If there’s not much data in a particular area, then the AI engines will make things up (“hallucinate”) rather than say “I don’t know.”
4. AI is a reflection of humanity. It only knows what we tell it. When the world’s knowledge is filled with conspiracy theories, pseudo-science and urban myths, that’s what it is most likely to play back to you.

Of course, there is a vast amount of investment going in to improving all of these points. One day, perhaps we’ll be able to trust AI engines on every topic, and they’ll be super efficient. For now, I strongly suggest to use with care!

Note: Copyright and intellectual property is a massive issue with AI, with many questions still unanswered. When the rules and regulations catch up, will it be too late to tell the AI to "forget" what it's already learnt? How do we identify where a particular source has been used? Should people explicitly give permission for their materials to be reused by an AI?

Impact on work and learning

Already things are changing for many, many people. I know it’s a cliché, but the closest parallel really is the industrial revolution. Although at a much faster rate.

The industrial revolution made a noticeable impact over the course of about 50 years from 1800 onwards (Clark G. 2014 Handbook of Economic Growth). The AI revolution is changing economies and labour markets far more quickly (see box).

Just in our small world of learning technology and elearning, already there are significant, long lasting things happening:

• Translation companies are struggling to compete with the growing capabilities of automatic translation engines. The era of the Babelfish is coming soon, if not already here.
• Companies like Fortell.ai are redefining how we measure impact of our initiatives, by conducting interviews at scale, in multiple languages, and performing automated analyses of the results.
• Every content development tool on the planet seems to be incorporating AI to rapidly produce even more elearning than ever before. Examples include Kwantic and Elucidat, among many others. Whether this is going to make a difference to the outcomes is debatable…
• Users within organisations are being given tools so they can query the knowledge in their organisations without wading through vast quantities of irrelevant search results. But see point 4, above - garbage in = garbage out. Examples include Gitbook and Fuse Universal, but there are many others.
• Teachers and lecturers are building AI into their assessment processes [Imperial college business school case study]
• Students are using AI engines to respond to assessments, leading to schools, colleges and universities rethinking their assessment strategies [Chartered College of Teaching Impact Article]

In my own work, I use AI in a small way - to rapidly generate ideas, to manipulate data, and to provide support as I learn new coding languages. It’s great at telling me where I’ve gone wrong, and explaining how to make things better. Some things, that would have taken hours, now take a matter of minutes.

None of this is threatening. AI is a new tool to learn how to use effectively. Some people will use it to create monstrosities, without thought or meaning. Whilst others will use it to enhance what they already do - allowing them to do more, and better.

Perhaps I’m being too optimistic? There obviously is a tangible threat to those whose work is being swallowed up by AI. Just as textile workers lost their jobs to the new factories. [WorldHistory.org]

But, as someone said recently, the genie is out of the bottle. So, now, the question is what do with it?

If you'd like to discuss this article, or how I can help you, get in touch.

Posted: 13 September 2024

Tags: Solution design Supplier selection Technology reviews

Related articles