Artificial intelligence (AI) comes with a wealth of possibilities – and many ethical concerns. Dr Amanda Williamson (University of Waikato and Deloitte) and Professor Nick Agar and Associate Professor Te Taka Keegan from the University of Waikato discuss some reasons to be wary of the new technology, the feasibility of laws regulating the industry and some suggestions for individuals navigating this new world.
Did you know?
The Cambridge Analytica data scandal in 2010 involved the harvesting of personal data without consent from millions of Facebook users. The data was used for political gain.
Questions for discussion
- How do you define ‘social cohesion’?
- How do you think the rise of social media has impacted social cohesion?
- What new jobs or new learning for current roles will be needed to regulate and safeguard against the abuse of AI?
- Nick suggests that AI and generative language models will be more transformative than social media. How do you think social media has transformed society? How do you think that AI and language models will transform society?
- Some of the speakers suggest ways to reduce the possible harms of AI and generative language models such as the ‘right to be forgotten’ when we request it. What regulations or brakes would you like to see applied?
Transcript
Dr Amanda Williamson
If we look back in history at the big scandals we’ve had with technology, for example, the Cambridge Analytica scandal, where our preferences were nudged in certain directions due to a technology that we didn’t even know existed most of us back then.
So my concern is that we don’t even know how AI will impact social cohesiveness. It’s very likely to have an impact that we cannot comprehend at this point in time.
Professor Nick Agar
Don’t look to me to reassure you, but I mean, how did social media affect social cohesiveness? In Mark Zuckerberg’s fantasy world, it brought us all closer together – not really what happened. So, yes, I think, generative AI – ask the same questions, be similarly suspicious.
Associate Professor Te Taka Keegan
In terms of social cohesiveness, I think it’s always important to meet people face to face. So I think technology has taken that away, and if we can find opportunities to still meet people face to face, be the audience in the room listening to live speakers. The Māori saying is kanohi ki te kanohi, rae ki te rae, tinana ki te tinana. So just face to face.
So I think one avenue, as well as what’s been said about having that ability to question what you’re seeing and what you’re hearing and what you’re believing, is to find avenues to meet face to face. I think that’s quite important.
Professor Bryony James
Is it possible for regulation to stay ahead of the exponential growth in the abilities of these models? In what ways do you think AI should or indeed could be legislated?
Associate Professor Te Taka Keegan
No. I don’t think it’s possible for regulation to stay ahead because we’re not sure of the impacts. It’s like when they first built the motor cars. Then they decided to – well let’s have speed limits and let’s drive on one side of the road. We’re not sure what are the impacts so we can’t build regulations cause we’re speculating.
So really we have to have a few accidents. We have to kill a few pedestrians. And never – OK, this isn’t a good thing – but I think that, once we understand some of the true impacts, then we should have the ability to regulate.
And like, for example, in terms of AI, one of the issues we have with Māori data sovereignty is finding out where it got its data. If we could regulate AI systems to show us where you got your data, I think that would put a lot of brakes on things. So there are avenues in terms of regulation, but we’re not quite sure yet what we need to regulate against.
Professor Nick Agar
It’s hard to regulate a concept like AI, but it’s definitely possible to regulate a company. And that’s, I think, the major threat. It’s what humans will do with these things. They’re the ones who are incentivised to create these texts. Well, maybe because they make money out of them. And if you think that social media made billions, I don’t know, this is much more transformative.
Are we going to see the world’s first trillionaires? Now, I think we have power over that if we choose to.
Dr Amanda Williamson
My biggest concern when it comes to regulation is the possibility to inhibit innovation in ways that we really need help. You know, we’ve got a big climate crisis right now and we need really smart tools to deal with that. But on the other hand, I think the scariest thing that we do need to regulate against is personal information. So right now, if you are using tools like ChatGPT and you are putting that information into the model, it will specifically tell you that that information will be used for the model’s training.
Which means that, if you work in an organisation and you’re putting human resource information or any proprietary information into the model, it can be sucked in there.
So I think that what we need in terms of regulation is to have the ability to be forgotten and to allow these models to forget us when we request it.
Acknowledgements
Dr Amanda Williamson, Senior Lecturer in Innovation and Strategy at the University of Waikato and Manager in AI & Data Consultancy at Deloitte.
Professor Nick Agar, Philosopher and Professor of Ethics, University of Waikato.
Associate Professor Te Taka Keegan, Department of Software Engineering, Co-Director of AI Institute Māori/Artificial Intelligence Institute and Associate Dean Māori for Te Wānanga Pūtaiao (Division of HECS), University of Waikato.
Professor Bryony James, Deputy Vice-Chancellor (Research), University of Waikato.
Video footage courtesy of the University of Waikato. The video is from a recording of the Kaupapa Kōrero panel discussion Bots vs Beings: How Will AI Impact Your Life and Work? held at the University in June 2023.
The full discussion with Professor Mike Duke, Dr Amanda Williamson, Professor Nick Agar and Associate Professor Te Taka Keegan and facilitated by Professor Bryony James can be viewed here.
