AI? Isn’t that ChatGPT? This was all of us once, stepping into the unknown world of Artificial Intelligence. In his talk at our annual event in March 2024, Dan Poulter, senior customer development manager from at JISC said “AI has moved very quickly over the past 12 months, and even more quickly during the last six”.
AI is evolving at pace and will undoubtedly form part of all our roles. A quick show of hands at our event found that almost all of our attendees had at least dipped their toes into the AI pond.
But what exactly is generative AI? What challenges could we face in using it? And what should we expect from the AI generation of employees?
Generative AI is much more than ChatGPT
For those of you who are new to AI, ChatGPT is natural language processing chatbot. According to a recent survey, ChatGPT is the most popular generative AI tool, receiving 14 billion visits between September 2022 and August 2023. AI tools and applications are developing at the speed of light. Whilst ChatGPT has been the most headline grabbing tool, it is far from the only one.
“Generative AI is a type of artificial intelligence technology that can produce various types of content, including text, imagery, audio and synthetic data.” ~ Techtarget.com
If you look at tools like Adobe photoshot, HubSpot and social media scheduling tools, generative AI integrations are there to greet us at every turn. Even posting on LinkedIn comes with the option to utilise AI.
In Dan’s talk he split generative AI platforms in six core categories;
- Chat: chatbots can hold a conversation with the user, these include ChatGPT and CoPilot.
- Search: unlike Google, given the correct prompts these tools can rapidly summarise content from across multiple sources in one seemingly concise response. However, we should note here that 47% of marketeers say that generative AI has provided them with inaccurate information.
- Images: AI image editing creation is on the rise, with integrations now within PhotoShop and Canva. However, this isn’t without its sometimes-hilarious challenges. It’s also important to mention here the issue of AI bias; influenced by wider societal bias, it’s key to ensure any output is aligned with your diversity and inclusion policies.
- Code: AI can now read images. Dan showed us an amazing example of Relume reading an image of a web page, and the tool writing the code to turn it into a webpage.
- Writing: One of the most popular uses of generative AI, it can of course be used for writing. But proceed with caution. As noted above, AI is often inaccurate. And if you’re aligning you copy to a tone of voice, any content will require additional work to ensure it doesn’t sound like every other article written using AI.
- Presentation creation: Finally, you can use AI to create presentations for you. Instead of using time formatting power point presentations, you can use AI tools such as Gamma to layout your presentation for you, leaving you more time to refine and focus on your content.
The opportunities are endless, but what are the copyright and IP implications?
Artificial intelligence relies on machine learning. And everything you type into ChatGPT feeds the beast. Samsung found itself in hot water in 2023 when employees used ChatGPT to assist the with tasks. Sensitive data including source code and internal meeting notes relating to their hardware were leaked on the platform. This prompted Samsung to ban all use of AI on their company owned devices.
Interestingly one of the attendees at our event said that he would discipline anyone in his team if he found them using ChatGPT. Many people in the room shared concerns about the copyright implications of using such platforms. What are the options for organisations wanting to harness the power of AI, without putting sensitive information at risk?
Dan says that it’s imperative for organisations to adopt a clear AI policy for all employees, ensuring policies are clearly communicated. Tools such as Microsoft CoPilot can be safely used without putting IP at risk. When working with partner organisations there is no harm in disclosing that you use AI, what your policies are, and how you’re keeping their information safe.
What should we expect from the next generation of AI savvy employees?
Jisc work heavily with colleges and universities, addressing the challenges and opportunities of AI in education. One of the key concerns raised by our attendees was the learning implications of AI for students. If AI can conduct research, summarise papers and write dissertations, will they have the skills required to enter the workplace?
As a founding member, we were joined by Shona Bettany, a Professor at the University of Huddersfield, who was able to address this challenge. Firstly, it’s important to acknowledge that AI is here to stay, and having a generation of AI literate employees may present opportunities for employers. She also talked about how AI is being managed at her university; course leaders can decide how much AI intervention they will allow, if any, based on three tiers. Students must disclose how and what AI tools they have used, which will be taken into consideration when marking. The key is to find a balance between embracing AI and ensuring positive learning outcomes for students.
Applying AI tools to your work:
AI presents opportunities for partner and ecosystem professionals to make efficiencies and streamline processes. Here are just some of the ways you could integrate AI into your work:
Proof reading: AI is incredibly useful for proofing long prose and suggesting changes. Tools: CoPilot and Grammarly.
- Research: If you’re in need of 10 fast facts on a particular topic, all you need is a suitable prompt. Tools: CoPilot, Gemini, Perlexity
- Code: Generate code quickly, with tools now reading hand drawn images. Tools: OpenAI, Gemini, GitHub CoPilot
- Content writing: If you really really need help writing content, proceed with caution and ensure you apply a human touch to the work that’s produced. Tools: Jasper, Writesonic, Grammarly
- Images: For rapid image editing, but expect the occasional interesting result. Tools: Midjourney, CoPilot, Stability.ai
- Chat: If you’re developing a brand voice or want to explore personas, experiment with prompts within chat.
- Tools: CoPilot, OpenAi, SnapChat, Claude AI is here to stay, and as ecosystem professionals, it’s within our interests to embrace and harness the possibilities it brings.
Find out more about the innovate ways in which JISC is utilising AI for its members and customers below.