after it allowed an altered video of# President Joe Biden to remain online.

This week, Meta's oversight board called the# company's manipulated media policies -- quote## -- "incoherent and confusing."

Meta now says# it's taking a number of measures to address## A.I.-generated content on its platforms, including# labeling them as such in the months to come.

Nick Clegg is president of global affairs at Meta,# and he joins me now to discuss these concerns.

Nick, welcome back to the# "NewsHour."

Thanks for joining us.

NICK CLEGG, President For Global# Affairs, Meta: Good to be AMNA NAWAZ When you begin to label images generated# an d Threads and Instagram, it's# going to look like this.

There's## going to be a little watermark in the# corner and a l I guess the question is, how do you know it's A.I.# if it's not generated on your own platfo have the tools to detect an A.I.-generated# image if it's created somewhere else?

NICK CLEGG: Yes, I mean, you make# exactly the right distinction,## which is, of course, if people use# our tools to generate A.I.

images,## we know that and we can watermark# that.

And we do that already.

So if you use one of our tools to concoct an# image using A.I., there's a wh ich makes it very clear that it's not a# sort of real photograph taken by a human## being.

Your question, quite right, is,# what happens when someone uses someone## else's tools -- and they're more and more# of them and they're multiplying -- seeks to share it with their friends and# post it on Instagram, Facebook, Threads?

And I think we have made some real progress as an# industry within the -- kind of under the umbrella## of an organization called Partnership for A.I.,# which brings together a lot of the main players## to develop common standards so that, when# we, as it's sort of called in the jargon,## ingest, in other words, when we bring on to our# platform because someone's sharing it on Instagram## or Facebook something which has been concocted# elsewhere, has been generated elsewhere, there is## an invisible watermark, something which allows# our technology to automatically identify it.

And then we can choose, as we will choose,# to label it, so users can distinguish between## what is human and not human, what is# synthetic and what is not synthetic.

AMNA NAWAZ: And to be clear, Nick -- I apologize## for the -- is that somethi NICK CLEGG: It's something we can# do now with images and we will be## instituting in the weeks and months to come.

But the difficulty is, the technology doesn't# ki nd of common standards on audio or video# content.

And, of course, there's always the## risk that people who want to use these tools for# malicious purposes will try and evade those rules.

I don't think it is a perfect solution.

But just## because it's not perfect does not# mean that we shouldn AMNA NAWAZ: Right.

now in keeping with the state of# the technology as AMNA NAWAZ: Well, let me ask you about your# policies as they exist today, because it was## your own oversight board that said your current# policies are There was this video that spread# on Facebook last year.

It was the## unaltered original video showed President# Biden placing an "I voted" stic granddaughter's chest.

The manipulated# one, the fake one made it look as if he## was repeatedly inappropriately touching# her.

And that spread quite a bit.

You made the decision to leave it up.

Why# did that video not violate your rules?

NICK CLEGG: Well, I should be clear.

The oversight# board said that we were right to keep it up,## and they actually criticized us for being# inconsistent about when we remove content.

And they basically said that trying to play# Whac-A-Mole and remove individual bits of## content is inconsistent and not effective.

So# what we're saying this week is very much in## keeping with the critique that the oversight# board delivered.

Namely, we are now going to## label much more visibly for users on a much wider# scale content that is artificially generated.

(CROSSTALK) that video showed a presidential# candidate doing something he did not do.

How worried are you that those kinds of# videos are going to be spread even more## widely, and they actually could help spread# dangerous information in a key election year?

NICK CLEGG: Well, I don't think# it will be that dangerous as## long as we can tell users that it is# An d do remember, of course,# that any synthetic content,## any A.I.-generated content about any of# the candidates will immediately be noticed## by those candidates.

So I don't think# there's much risk that it'll go somehow## unnoticed and many millions of people# will see it and no one will real And we will then, of course, be able to move -- if# our automated systems in the way that I described## earlier haven't caught it already, we will then# be able to label it.

So I'm -- I certainly don't## want to be complacent, but I think considerable# efforts will be made by us and others to make## sure that users can, to the very best of our# powers, be able to distinguish between something## that has actually been produced authentically# and something which has been produced by A.I.

And I hope also, in -- alongside our policies,## which mean that no one can run a political ad# on Facebook or Instagram without declaring an disclosing that they have used A.I.

-- and# if they do that, and they systematically## try and avoid that policy, we will,# of course, take action against them.

(CROSSTALK) but I want to make sure I'm# respectful of your NICK CLEGG: Sure.

AMNA NAWAZ what does that mean?

You wil NICK CLEGG: Yes, we have got# a whole range of penalties,## which, of course, could start with a# right through to basically saying they# can't run ads on Faceboo There's a range of penalties that we# can apply.

And I would have thought most## mainstream campaigns would want to retain# the ability to communicate with political## ads on our platform.

So I think they have# got a huge incentive to play by our rules.

AMNA NAWAZ: And, Nick, what's the# role for lawmakers here?

I mean,## there was a bipartisan bill that would# officially ban that deceptively portrays candidates# for federal office in political ads.

Do you support that bill?

NICK CLEGG: I haven't looked# at that bill specifically,## but I do think there is very much# At the end of the day, we're talking here# about the policies that we have developed,## to the best of our abilities, about how# users will be able to and viewers will be## able to decide whether -- for themselves# whether something is A.I.-generated or not.

But when it comes to elections, the elections# belong to the people, to the country,## not to big tech companies.

And it really# should be as part of the democratic process,## for democratic politicians to come# together across party lines and## decide for themselves the guardrails# that they want to apply in elections.

And we, of course, will abide by# whatever guardrails they come up with.

AMNA NAWAZ: That is Nick Clegg, president of# global affairs at Meta, joining us tonight.

Nick, thank you very much.

Good to see you.

NICK CLEGG: Thank you.