AI: Understanding the implications of fast developing technology

It’s hard to believe it’s only been a year since ChatGPT launched to the public. While AI tools have been around for many years, so-called LLMs (Large Language Models) – and ChatGPT in particular – have been instrumental in bringing them to the attention of people outside the tech industry. Increasingly these tools, called Generative AI are used in a range of ways to save time, to generate ideas and to shortcut research or analysis tasks. 

Previously, to use an AI tool you needed some coding knowledge or to be comfortable using complex prompts. Now popular tools like ChatGPT, Google Bard and image-generating tool Dall-E all respond to conversational language.  

There has inevitably been some alarm about the adoption of these tools and how they might replace humans to perform certain tasks. For the time being, this is most likely to be in customer support roles in the form of chatbots, data entry and analysis and anything involving routine or repetitive admin tasks. 

Concerns aside, these tools can be extremely helpful, particularly if you’re in a marketing, audience development or fundraising role, so let’s consider how you can use them now to support your day-to-day work.

Generating ideas

One of the most common uses of AI tools is to support content creation. There is criticism that content generated this way can be bland and generic – that's fair comment. The trick is to use them to save you time but not to replace writing that needs an authentic voice, like a blog post. 

You might ask the tool to come up with ideas but then write the copy yourself to ensure it aligns with your organisational tone of voice. For example, if you are responsible for social media adverts, an AI tool can help you generate different versions. Tell it who you are aiming at, what you want to achieve and even the style of writing. You can then ask it to refine its suggestions, to adapt the language and replace phrases you don’t like. 

If you then run an A/B test on the adverts, you can tell the AI tool which was more successful and, over time, it will learn what works best for your audiences and the quality of its output will improve. Equally, you can ask it to generate subject lines for your newsletters, trying different styles and prioritising different messages.

Considering data analysis

Another great use is data analysis. If, for example, you’re running Google adwords campaigns, an AI tool can analyse the effectiveness of certain keywords and identify trends in click through rates. (Note, if you’re using a free version such as ChatGPT 3.5, you will be limited in the amount of data you can post into the model for a single interaction. But if you’re new to the tools, it’s a good way to get comfortable with what is possible.)

Considering data analysis, one of the first data sets you might want to try is survey data. AI tools can summarise survey findings, extract key themes from text responses and run statistical analysis to identify trends. 

All this can save time, but you need to be mindful of privacy issues. You should never post sensitive or personally identifiable data into AI platforms. As a side note, with more people experimenting with these tools, it is sensible to develop an AI policy for your organisation. 

A great example of this is the government advice to civil servants which encourages staff using generative AI to consider the three ‘hows’. These are: How your question will be used by the system; How answers can mislead; and How generative AI operates. It also provides examples of appropriate and inappropriate uses.

General tips

Language generation or analysis are not the only areas where AI applies. Image generation tools, such as Dall-E can also be good for marketers where, for example, you’re need an image to accompany a blog, or perhaps when you need something non-specific, but that adds colour and interest. 

There are lots of available resources to help you get the best out of these tools but, the more you use them, the more skilled you’ll become in understanding what generates the most successful outputs.

Here are some general tips for using AI tools in your day-to-day work:

●    Be very clear in how you describe what you want the tool to do. The clearer you are, the better the output.
●    If you are using a tool for data analysis, the accuracy of your data matters. The quality of input data directly impacts the quality of the output.
●    Work with the team to consider how tools might be of use to your particular organisation – play, experiment, have fun. One criticism of tools like ChatGPT is they sometimes ‘hallucinate’ or make things up. This is serious when dealing with factual information but the flipside is they can help generate ideas. Perhaps think of them as a creative assistant and get them to come up with ideas that you subsequently refine.
●    Set organisational standards for how you will and won’t use these tools – both quality standards and ethical standards.
●    Keep up to date on the latest developments. For example, ChatGPT recently announced that it can hear, see and speak. This is likely to have a significant impact on the adoption of these tools and their impact on our lives.

There is no doubt we need to understand the implications of these tools as they develop at such fast pace. What we use will be materially different in 12 months. How we regulate AI is likely to be a key inflection point in our history. For now, used ethically, Generative AI tools can, at a minimum, help us save time, generate ideas and shortcut common processes. 

Katie Moffat is Director of Sector Strategy at Substrakt.
 substrakt.com/
@substrakt

This article is part of a series contributed by Substrakt exploring the many ways in which arts and cultural organisations can embrace the world of digital.