My Lords, as is so often the case on these issues, it is daunting to follow the noble Baroness as she has addressed the issues so comprehensively. I speak in support of Amendment 57, to which I have added my name, and register my support for my noble friend Lord Holmes’s Amendment 59A, but I will begin by talking about the Clause 14 stand part notice.
Unfortunately, I was not able to stay for the end of our previous Committee session so I missed the last group on automated decision-making; I apologise if
I cover ground that the Committee has already covered. It is important to start by saying clearly that I am in favour of automated decision-making and the benefits that it will bring to society in the round. I see from all the nodding heads that we are all in the same place—interestingly, my Whip is shaking his head. We are trying to make sure that automated decision-making is a force for good and to recognise that anything involving human beings—even automated decision-making does, because human beings create it—has the potential for harm as well. Creating the right guard-rails is really important.
Like the noble Baroness, Lady Kidron, until I understood the Bill a bit better, I mistakenly thought that the Government’s position was not to regulate AI. But that is exactly what we are doing in the Bill, in the sense that we are loosening regulation and the ability to make use of automated decision-making. While that may be the right answer, I do not think we have thought about it in enough depth or scrutinised it in enough detail. There are so few of us here; I do not think we quite realise the scale of the impact of this Bill and this clause.
I too feel that the clause should be removed from the Bill—not because it might not ultimately be the right answer but because this is something that society needs to debate fully and comprehensively, rather than it sneaking into a Bill that not enough people, either in this House or the other place, have really scrutinised.
I assume I am going to lose that argument, so I will briefly talk about Amendment 57. Even if the Government remain firm that there is “nothing to see here” in Clause 14, we know that automated decision-making can do irreparable harm to children. Any of us who has worked on child internet safety—most of us have worked on it for at least a decade—regret that we failed to get in greater protections earlier. We know of the harm done to children because there have not been the right guard-rails in the digital world. We must have debated together for hours and hours why the harms in the algorithms of social media were not expressly set out in the Online Safety Act. This is the same debate.
It is really clear to me that it should not be possible to amend the use of automated decision-making to in any way reduce protections for children. Those protections have been hard fought and ensure a higher bar for children’s data. This is a classic example of where the Bill reduces that, unless we are absolutely explicit. If we are unable to persuade the Government to remove Clause 14, it is essential that the Bill is explicit that the Secretary of State does not have the power to reduce data protection for children.