Malcolm Byrne seeks safeguards on AI chatbots aimed at children
Malcolm Byrne has sought leave to introduce the Artificial Intelligence Companion Services Protection of Children Bill 2026. He says the Bill would restrict AI companion services for children and strengthen safeguards against chatbots that imitate human relationships.
We publish thousands of recordings to make Irish politics transparent and resistant to manipulation. Spotted an error? Report it — together we are building a reliable archive of Irish politics.
Other speeches
Tego samego dnia All speeches from this day →
Charles Ward
Charles Ward: Donegal faces higher fares and poor bus services
Darragh O'Brien
Darragh O'Brien: Fare rises will fund expanded services
Paul Murphy
Paul Murphy challenges data centre electricity costs
Gerard P. Craughwell
Gerard P. Craughwell says Ireland is underfunding defence
Martin Daly
Martin Daly: Cash is a vital safety net
Carol Nolan
Carol Nolan: Rural Ireland Shouldn’t Pay More for Less
Transcript
I'm happy to seek leave to introduce the Artificial Intelligence Companion Services Protection
of Children Bill 2026.
I certainly believe that this should be the best and safest country in the world in which
to grow up as a child, and I think that principle should apply to all that we do.
And I think colleagues in here will know that I am an evangelist for the use of technology
and indeed artificial intelligence, where it can be used responsibly to solve many of
our challenges.
And certainly we can use technology to help children in the area of education, and indeed
with parental safeguards, technology can be used in the area of children's play.
I have my doubts as well about the efficacy, for instance, of a social media ban for those
under the age of 16.
But what this legislation seeks to address is the challenge of the growth of chatbots,
particularly targeted at children, and especially where there are no safeguards in place.
And I think we definitely need to be quite worried, and even with reference to the last
question that was raised, about the rise of the anthropomorphic chatbot, the chatbot that
has human characteristics, and where certainly the child will feel that he or she is interacting
with a human.
There are enormous challenges that we are now seeing with chatbots being developed targeted
specifically at children.
Children are being found to develop attachments and emotional bonds, and in many ways that
they're finding that they can engage more regularly or more readily with these chatbots
as opposed to with their human peers.
And there's obviously from that the dangers in terms of the skills that we all take for
granted in terms of empathy, human emotion, that that will be compromised.
But there's also the danger of inaccurate advice that can come from some of these chatbots,
leading children to have destructive thoughts, to have an impact on their mental health,
and obviously also on loneliness.
And these are not some, you know, people kind of think about, oh, the robots are coming,
and this is abstract.
We're talking already in place about quite a number of chatbots, particularly anthropomorphic
chatbots, being aimed at children.
When I was preparing this bill, I hadn't anticipated that there's now a proposal at EU level under
the EU Kids Act, which will in fact look to address some of these measures.
And what is critical is that we look at when some of these products are rolled out, that
the principle of safety by design is embedded within this.
That what we do is we look to a situation whereby we prevent an emotional dependency
by children on these AI devices.
But what I would say is if the EU Kids Act doesn't proceed quickly enough, that we should
move here in this state to try to address it.
And I'm very conscious that at present the US Federal Trade Commission is already investigating
the interaction and monetization of chatbots that mimic human conversation and emotion.
I think in terms of the EU Kids Act, and the principles as well that underpin this, is
that you cannot use the design of a chatbot to simulate a human relationship with a child.
And I think we've got to build that in to our own legislation.
So what this bill is proposing is establishing a prohibition on the making available of specific
AI companion services and intimate AI friendship and companion services to children.
It will empower Commissioner Mian to supervise and enforce compliance, and indeed it will
amend the Online Safety and Media Regulation Act with regard to the provision of these
services.
And in particular, where we're going to be talking about the online safety codes, that
these will apply to cover any of these chatbots, particularly that are aimed at children.
I'm very conscious about how quickly the technology is changing, and it's very difficult for us
as legislators to keep up.
And I think the government is moving in a way as quickly as possible, but in this specific
area around the protection of children, it is vital that we now take action.
And in that regard, Whip, I would ask that there's always a challenge in terms of for
backbenchers introducing private members' bills to get them progressing beyond first
stage of debate.
So I would ask that you might consider the Business Committee to look at a provision
on, for instance, on Thursdays, that a number of private members' bills could be considered.
We sort of have statements on all sorts of issues during the week, whereas there's limited
enough time to allow for further debate on legislation, which I believe on particularly
important issues like this we should have.