NEW YORK, Sept 18 (Reuters) - U.S. lawmakers questioned

tech executives on Wednesday about their preparations for

battling foreign disinformation threats ahead of elections in

November, with both the senators and executives identifying the

48 hours around Election Day as the most vulnerable time.

"There is a potential moment of peril ahead. Today we are 48

days away from the election... the most perilous moment will

come, I think, 48 hours before the election," Microsoft ( MSFT )

President Brad Smith testified at the hearing, held by the U.S.

Senate Intelligence Committee.

Senator Mark Warner, who chairs the panel, agreed with Smith

but said the 48 hours after the polls close on Nov. 5 could be

"equally if not more significant," especially if the election is

close.

Policy executives from Google and Meta,

which owns Facebook, Instagram and WhatsApp, also testified at

the hearing.

Elon Musk's X was invited to testify but declined, several

senators said. An X spokesperson said the reason was that the

company's invited witness, former head of global affairs Nick

Pickles, had resigned earlier this month.

TikTok was not invited to participate, according to a

company spokesperson.

To illustrate his concern about the time immediately before

people vote, Smith referred to a case from Slovakia's 2023

election, in which a purported voice recording of a party leader

talking about rigging the vote emerged shortly before the

election and spread online. The recording was fake.

Warner and other senators also pointed to tactics revealed

in a U.S. crackdown on alleged Russian influence efforts earlier

this month, involving fake websites made to look like real U.S.

news organizations including Fox News and the Washington Post.

"How does this get through? How do we know how extensive

this is?" Warner asked the executives. He requested that the

companies share data with the committee by next week showing how

many Americans viewed the content and how many advertisements

ran to promote it.

Tech companies have largely embraced labeling and

watermarking to address risks posed by new generative artificial

intelligence technologies, which have made fake but

realistic-seeming images, audio and video easy to produce and

raised concerns about their impact on elections.

Asked how the companies would react if such a deepfake of a

political candidate were to surface immediately before the

elections, Smith and Meta's President of Global Affairs Nick

Clegg both said their companies would apply labels to the

content.

Clegg said Meta also potentially would suppress its

circulation.