The rate of change in AI is currently rapid and much of our current understanding, including the tools and their use cases, will be outdated in a matter of weeks. Although it is impossible to accurately predict where the technology is going, it is reasonable to assume AI will be able to generate more realistic and diverse content, such as images, videos, texts, sounds and even code. This may enable new forms of creativity, entertainment, education and communication. However, it will also pose ever-greater ethical challenges in areas such as deepfakes and plagiarism.
What kinds of AI tools are out there?
There has been an explosion of AI tools in recent months, and therefore providing an exhaustive list would be impossible. However, these are some of the most common:
LLM Chatbots
An LLM (‘large language model’) chatbot is an AI-driven tool that understands and generates human-like text. Users interact with LLM chatbots through natural language prompts, and they deliver context-sensitive and often coherent responses. Some examples of these tools are as follows:
Writing assistance tools
AI writing tools are software programs that use artificial intelligence (AI) and natural language processing (NLP) to help users improve their writing skills. These tools can help users with various aspects of writing, including grammar and spelling checking, sentence structure analysis, and even content generation. Some examples include:
AI Image generators
AI image generators are tools that can turn text-based prompts into context-driven visuals. They utilise training on image datasets to generate a wide array of visuals, ranging from photorealistic scenes to artistic interpretations. Examples include:
Captioning tools
meetings. They can enhance accessibility, comprehension, and engagement by providing real-time captions for individuals:
Speech generators
AI speech generators are can transform written text into natural-sounding speech. These generators find applications in various domains, including audiobook narration, voice assistants, and content accessibility. Due to their ability to quickly and accurately clone individual voices, there has been concern over the potential for these tools to for misinformation and deepfakes (see below). Some examples include:
Programming/software development AI tools
AI software development tools can refine and expedite the software development process. By aiding developers with tasks like code generation, error identification, and optimisation, these tools have the potential to speed up the development process.
- GitHub Copilot
- ChatGPT (although its primary use is as an LLM chatbot (see above), ChatGPT can be used as a programming/development assistant)
Concerns around usage of AI tools
Bias
Most AI models have undergone training using data sourced from the internet. However, the internet is not a neutral or unbiased environment, leading AI tools to inadvertently mirror these biases in their outputs. While developers will continue to attempt to address this issue, the complexity and pervasiveness of biases make it difficult to see how this can ever be resolved.
Misinformation
AI tools are not able to tell if the information it has been trained on is true or not, and they will not be able to guarantee that their generated content is accurate. As a result, misinformation is common and can be presented with the same degree of confidence as the truth. Therefore, it is important to approach AI-generated content with a critical eye and double-check the facts using trusted sources.
How are these tools being used?
One of the better ways to consider AI tools is as genuinely useful assistants that augment human capabilities. This augmentation saves time, which means that anyone using it should have more capacity for other things. The opportunities this opens will likely depend on the individual and the kinds of activities that are valued within the institution/faculties/teams they work within.
Below is a list of the kinds of applications that these tools can be used for:
LLM Chatbots:
Providing academic and administrative support to students, such as answering course-related questions or assisting with enrolment processes.
Offering personalised career guidance to students, such as recommending job opportunities based on their skills and interests.
Writing assistant tools:
Enhancing student writing skills by providing real-time feedback on grammar, syntax, and clarity.
Streamlining the process of creating academic content, such as research papers or dissertations, by offering suggestions on organization and formatting.
Image generators:
Providing visual aids for academic presentations, such as graphs or diagrams, based on text-based input.
Creating customised visual content for marketing materials, such as social media posts or advertisements.
Captioning tools:
Making video content accessible to all students, including those with hearing impairments, by generating accurate closed captions.
Assisting with note-taking during lectures by providing real-time transcriptions of the spoken content.
Speech generators:
Offering a new way to create educational content, such as podcasts or lectures, by generating synthetic speech from text input.
Enhancing the accessibility of audio content, such as audiobooks or podcasts, by creating a synthetic voice for those who are unable to record their own.
Programming/software development AI tools:
Assisting with the testing and debugging of code, such as by identifying common errors or suggesting improvements to code structure.
Streamlining the process of software development by automating routine tasks, such as generating boilerplate code or creating documentation.
Article authors
David Ashworth & Martin Jenkins
