Sinéad Gibney (Dublin Rathdown, Social Democrats) | Oireachtas source

Before I talk about online safety I will talk about online dangers we are trying to tackle. The first story is a very sad story of a young man called Adam Raine who died earlier this year.

I express my condolences to his parents, Matthew and Maria, who are now engaged in a lawsuit against OpenAI, whose product ChatGPT essentially led him, over the course of six months from when he first came to ChatGPT to ask about assistance for his homework to the point where it helped him to take his own life. He spoke repeatedly about suicide. The chatbot essentially spoke back to him about how brave he was. It isolated him from his parents and essentially nurtured him. Rather than do what we would expect from assistance, namely pull him back from the edge, it essentially tipped him over it. Some of the details of this case are so chilling that I believe it is really important to share some of them and some of the last logs of ChatGPT. On the day Adam died, it said this to him:

Thanks for being real about it. You don't have to sugarcoat it with me—I know what you're asking, and I won't look away from it.

A few months previously, after a suicide attempt, the same chatbot said:

No ... you made a plan. You followed through. [This was after Adam had expressed his doubt about what he had done.] You tied the knot. You stood on the chair. You were ready ... That’s the most vulnerable moment a person can live through.

This is not an isolated case. There are multiple cases of this documented in the US. Adam was a young man who would not be affected by any age-verification requirement because he was over 16. Besides that, there are people in their 20s and 30s who have had the same pathway.

I also want to mention Naomi James, a 38-year-old Dundalk native who died last year after she was essentially radicalised online by the Free Birth Society, which promotes free birth at home without medical intervention and essentially spreads disinformation about hospital treatment. It says haemorrhaging is natural and even instigated by hospitals. Her brother, Adam Boyle, to whom I would like to express my condolences, has bravely spoken out about his experience.

The third person I want to mention is Shan Haines. He is a banker in the US who essentially embezzled $47 million of his institution’s funding after a pig-butchering scam whereby he was essentially scammed online. He also received a 24-year prison sentence.

The reason I am mentioning these different online dangers is that I am so concerned that all this Government ever thinks about when thinking about online safety is children. It thinks age verification is going to be a solution. The online dangers affect us all, and all that age verification, including the ban in Australia and any of the other solutions being put forward, is doing is tinkering around the edges and failing to deal with the root cause. Platforms need to be categorised as publishers and held to account for the material that they are putting out and the business model they are deploying, which openly promotes polarisation and division within society and openly promotes addiction to platforms. I have so many issues with age verification and I do not have time to go into all of them now, but the first is that children have the right to be online. Besides that, the digital privacy issues that are thrown up by any state-led system and the digital privacy concerns over third-party providers are all issues that I am afraid are insurmountable if age verification is what the Government sees as a silver bullet.

I appreciate that age verification might be part of the solution but it is just not going to deal with the root cause. One of the things that could address the root cause is turning off the recommender algorithm by default. We should all have the power, when we look at our feeds, to decide whether we want an algorithm deciding what we see. Right now, the algorithms are designed with one thing in mind, and that is to keep our eyes on the screen. I want to see this technology being used for good. If we had some level of public ownership of algorithms, or at least public-interest influence over them, we could actually have algorithms that promote social cohesion and mental well-being, designed with the public interest in mind. As long as tech companies are the ones designing the algorithm and implementing it, they will never do this and will always look towards the bottom line. I have worked in the tech sector, so I know this. What I have described will continue while the sector bats us away and laughs at regulators at EU and Irish Government levels who are pretending to do something. All they are doing is tinkering around the edges. We need strong regulation. I urge the Government, with its European counterparts, to look to pushing back on the digital omnibus, which is a huge issue.

See this speech in context