Experimenting with generative AI

Over the past couple of months, the Digital Innovation Lab in IT Services has been experimenting with different generative AI tools. Several blog-style articles were produced by providing specific prompts to a range of different generative AI tools designed to explore the possibilities, challenges and benefits of generative AI tools. Topics included: 

The content of these articles, which was generated from several different sources, has been left purposefully unedited by humans so that the output can be analysed and discussed. 

Our observations

We've set out some of our key findings about the impact of these tools for content generation at the University. We have used Bing Chat to help generate this article, but this time we have edited it slightly to ensure it reflects our findings accurately.  

Unleashing creativity 

Generative AI tools have shown immense potential in fostering creativity. By providing specific prompts, the tools were able to generate content on a wide array of topics. This not only saved time but also offered a fresh perspective on familiar subjects. For instance, the article on ‘Christmas and New Year traditions around the world’ provided a unique, AI-driven insight into global cultures. 

Enhancing accessibility 

The use of AI tools could significantly enhance accessibility, especially in the realm of immersive technologies. AI can create personalised experiences tailored to the needs of individuals with different abilities. For instance, AI algorithms can generate real-time captions or modify the visual elements to improve immersive experiences for those with hearing difficulties or visual impairments. Furthermore, AI can enable voice commands and gesture recognition, allowing users to interact with the virtual environment in a more intuitive and accessible manner. These are just a few ways that AI has the potential to make immersive technologies more inclusive, breaking down barriers and creating a more equitable digital landscape. 

Cybersecurity and AI 

One of the key areas where generative AI tools could have a substantial impact is in the field of cybersecurity. The ability of these tools to simulate potential cybersecurity scenarios in higher education has provided valuable insights, helping to strengthen defences and prepare for potential threats. 

The double-edged sword 

While generative AI tools offer numerous opportunities, they also present certain challenges. The quality of the generated content can vary, and there is always the risk of the AI producing misleading or inaccurate information. By leaving the AI generated content of the five articles unedited, the Digital Innovation Lab has provided a transparent view of the capabilities and limitations of these tools. No matter how content is created, it is always important to double check the facts and sources. 

Conclusion 

As we continue to explore the potential of these tools, it is crucial to navigate their challenges effectively to harness their full potential. The journey so far has been promising, and the future of generative AI at Oxford University looks bright.  

Further information about generative AI

New government guidance on the use of generative AI has recently been published which is worth a look through to find advice on responsibility, good practice and safety. 

You can also find lots of useful information and guidance about AI at Oxford in the ‘Related links’ section on the right of this page and by visiting the AI and academic practice page on the Centre for Teaching and Learning website.  

For further information about digital innovation at Oxford University, please contact innovation@it.ox.ac.uk.   

Got a view? Get involved! 

Explore the possibilities, challenges and benefits of generative AI tools and inspire others to learn about this new technology by joining the discussion in the Generative AI SIG Teams channel.  

Can you suggest other tools to use or other prompts? What would you like to ask AI?  

Get involved!