6 March 2023
Heading MBS presents the Q&A on AI with Prof. Johan Steyn (part two of two)
In this, the second half of a two-part series, our guest Prof. Johan Steyn answers the burning questions from the discussion Milpark Business School hosted titled, "ChatGPT & AI: Our creations have become creators".
Prof. Steyn is a Human-Centered Artificial Intelligence advocate and thought leader; and one of the top 50 global voices on AI as recognised by SwissCognitive.
Q: I'm just worried about plagiarism; how will it be detected?
There are a number of new platforms that claim to detect AI-generated text. In my experience, they are all fairly ineffective in detecting it. I guess plagiarism detection models will get better over time.
Q: How important is professional scepticism when dealing with generative AI output?
Professional scepticism is an important trait in any field that involves analysis, interpretation, or decision-making. When dealing with generative AI output, professional scepticism becomes even more important because of the potential for bias, error, or unintended consequences.
Generative AI is designed to generate new content, based on patterns and algorithms derived from large data sets. While these algorithms can be very effective in creating realistic and compelling content, they are not perfect, and may produce results that are flawed, biased, or inaccurate.
When dealing with generative AI output, this means taking a critical and analytical approach to the results, and being willing to scrutinise and validate the output using other sources and methods.
Q: If you ask it to apply Roman Dutch law, will it be correct?
As an AI language model, Chat GPT has been trained on a vast corpus of text data, and has a general understanding of legal concepts and terminology. However, it's important to note that AI models like Chat GPT are not legal experts and may not have a full understanding of the nuances and complexities of specific legal systems, such as Roman Dutch law.
While Chat GPT may be able to provide some general information or insights on Roman Dutch law based on its training data, it should not be relied upon as a substitute for professional legal advice or consultation. If you have specific questions or concerns about Roman Dutch law, it's always best to seek guidance from a qualified legal professional who has experience and expertise in this area.
Q: As a qualitative researcher, would you suggest using ChatGPT to enhance the clarity and coherence of participants' responses during transcription?
ChatGPT can be a useful tool for generating and clarifying responses during transcription, particularly in cases where participants may have difficulty expressing themselves clearly or coherently. However, it's important to use Chat GPT judiciously and to recognise its limitations. It may not always generate responses that accurately reflect the intended meaning or tone of the original statement, particularly in cases where the language used is highly technical or domain specific.
It is important to recognise that the use of AI tools like ChatGPT in qualitative research may raise ethical and methodological concerns, particularly with regard to issues of data privacy, transparency, and reliability.
Q: Can the AI (ChatGPT) include references in its write-up so we can read up further on the information it is providing?
If ChatGPT generates a response that discusses a particular concept or idea, it can provide a hyperlink to an external source or resource that provides more in-depth information about that concept or idea. This can be a useful way to provide additional context and support for the information being presented, and can help to enhance the reader's understanding and engagement with the topic.
It's worth noting that the ability of ChatGPT to reference external sources will depend on the specific use case and the available information on the topic being discussed. If the information being referenced is highly technical or domain specific, it may be more difficult for ChatGPT to provide accurate and useful links to external sources.
Q: I imagine that in the future we will be so dependent on this this that an AI crash of some sort may leave humans totally disabled. What measures do they have in place for this eventuality?
While an AI crash or failure is a possibility, there are measures that can be taken to mitigate the risks and minimise the potential impact. By developing backup systems and redundancies, designing AI systems with error detection and correction mechanisms, and establishing contingency plans and emergency protocols, we can help to ensure that our reliance on AI does not leave us vulnerable to catastrophic failures. It is up to governments and large corporates we rely on to have these measures in place.
Q: Must we wait for something to go wrong first before we start thinking about how to regulate the potential abuse that may arise from this technology... is it possible to be proactive with law-making in the future?
Sadly, sufficient regulations often only arise after a large-scale problem. There are a few good global initiatives to regulate this technology, but almost nothing has been done in SA so far. We should be proactive in law-making and it will require good collaboration between the government, academia, civil society and business.
Q: There was an article, UN says AI poses 'serious risk' for human rights. Please comment about this, Johan.
The United Nations has raised concerns about the impact of AI on human rights, citing the potential for AI to perpetuate or even exacerbate existing biases and discrimination. In a report issued by the UN's human rights office in 2021, the organisation warned that AI could pose a "serious risk" to human rights in a variety of areas, including privacy, freedom of expression, and equal treatment under the law.
The report also raises concerns about the potential for AI to erode privacy and freedom of expression, particularly in cases where AI is used for surveillance or content moderation. There is a risk that AI systems could be used to monitor individuals or groups without their knowledge or consent, or to censor or restrict access to information in ways that violate freedom of expression.
Q: How do we remain relevant in the face of this massive disruption?
The rapid pace of technological change and disruption is posing significant challenges to individuals and organisations across industries and sectors. To remain relevant in the face of this disruption, it's important to adopt a mindset of continuous learning and adaptation, and to prioritise the development of new skills and competencies that are in demand in the new technological landscape.
One key strategy is to invest in ongoing education and professional development, and to seek out opportunities to learn new skills and acquire new knowledge. This might involve pursuing additional formal education or training, attending conferences and events, or engaging in online learning or self-directed study. By remaining curious and open to new ideas and approaches, individuals can better position themselves to adapt to changing technological demands and stay relevant in the face of disruption.
Q: There's also the consideration of Third World countries getting left behind. Take for instance SA, how can we possibly take advantage of these technologies, still battling with load shedding?
You raise an important point regarding the potential for developing countries to be left behind in the face of technological disruption. While the adoption of new technologies can present significant challenges for countries that are still grappling with basic infrastructure issues like access to reliable electricity, there are strategies that can be employed to help bridge this gap and promote greater access to technology and its benefits.
One key approach is to focus on building the necessary infrastructure to support the adoption and use of new technologies. This might involve investing in the development of reliable energy systems, such as renewable energy, to help address issues like load shedding and provide a more stable and consistent source of power. It could also involve investing in broadband infrastructure and other forms of digital connectivity, to help ensure that individuals and organisations have access to the digital tools and resources they need to take advantage of new technologies.
Another important strategy is to focus on building local capacity and expertise, and developing programs and initiatives to help individuals and organisations build the skills and competencies they need to take advantage of new technologies. This might involve investing in education and training programs, developing mentorship and coaching programs, or providing access to resources and support networks to help individuals and organisations build their capacity to adopt and use new technologies effectively.
Catch up on the full webinar here: https://lnkd.in/df5t-3nU