A strange phone call awakened US officials to the role of artificial intelligence in the elections – the moment of greatest danger will arrive exactly one week from now | Foreign countries

A strange phone call awakened US officials to the role

As Democrats prepared to vote in their party’s primary in January in New Hampshire, many received a strange phone call.

Joe Biden a voice seemed to urge them to skip the vote.

– It is important that you save your vote until November, the voice said.

It explained that voting in the primary would only benefit Republicans Donald Trump’s in the selection.

It was a mass call message, i.e. Robo call, which was directed to the voters based on a previously collected playlist.

However, the voice was not the real Biden, it was a deep fake. It means an image, sound or video created by artificial intelligence that shows a person saying or doing something that they did not actually do.

The case is an example of how different actors are trying to influence the US presidential election with the help of artificial intelligence.

– We may not know until after the elections what the biggest threat of artificial intelligence to the elections is. So you have to be vigilant, says the professor Emily Bell from New York’s Columbia University in an interview with .

Bell is the director of the university’s digital journalism research unit.

The United States is politically divided. Artificial intelligence can be used to promote political goals.

In the case of New Hampshire, it was eventually revealed that the voice message came from the Democratic camp. In any case, the case alarmed the US authorities. The same could also happen on election day in November.

Legislation is lagging

Because of the New Hampshire case United States Communications Agency banned, under threat of a fine, mass call messages in which voices have been created with the help of artificial intelligence. The ban is based on consumer protection.

Otherwise, legislation has not kept up with artificial intelligence, although both parties are concerned about the use of deep fakes made with the help of artificial intelligence in elections.

In Congress, the parties have made three joint initiatives that aim to regulate the use of artificial intelligence in elections. None of them will be made into federal law before the election.

It is difficult for regulation to keep up with artificial intelligence despite the will.

– Yes, it says a lot that no law of any kind has been enacted, says Emily Bell.

The states have succeeded a little better in legislative work. In some of them, the use of deep forgeries in elections is prohibited.

Non-governmental organization of Public Citizen however, based on tracking, many states will go to the polls without such legislation.

Good evil AI

Artificial intelligence is also involved in the presidential election as a tool of the election authorities.

At best, it makes it easier to manage information about millions of voters, as election officials maintain voter lists, verify postal votes, and draw electoral district boundaries.

Licentiate in law, doctoral student Sarah Bender has studied the use of artificial intelligence in elections.

– Relatively little attention has been paid to how artificial intelligence affects voting rights issues, Bender writes In Michigan Law Review in a published thesis.

Bender states in his thesis that artificial intelligence has been found, for example, to recognize names of foreign origin worse, so that minorities are more often the target of errors.

Names that closely resemble each other may get confused, and in the worst case, the authorities will remove the person from the voter list.

It takes a lot from a voter to claim his rights if there is an attempt to prevent him from voting on election day.

In verifying the signatures of postal votes, there is also a fear that artificial intelligence will recognize other languages ​​worse than English and lead to unnecessary votes being rejected.

US politicians have also done various artificial intelligence experiments. For example, the mayor of New York Eric Adams has used artificial intelligence in his announcements, in which he sounds like he’s speaking Mandarin and numerous other languages ​​he doesn’t actually know.

The problem is that Adams did not immediately reveal to the public that the material was created with the help of artificial intelligence.

The moments of danger are only in November

When the owner of the messaging service X Elon Musk share an image made with the help of artificial intelligence Terrible about Harris in the uniform of a communist dictator, it is unlikely to affect the election result.

Rather, experts pay attention to the dangers of artificial intelligence on election day itself and after election day.

An easy-to-implement danger on election day would be a mass phone call in different languages ​​claiming, for example, that the voting period has been extended.

Such a privately targeted message would be impossible to catch and correct in public.

– In the United States, the election can be decided by a very small margin. It’s enough that only a small number get wrong information, Emily Bell reminds.

In the counting of votes, the danger lies with the election authorities. In the United States, up to half of them are politically committed and may have to act under pressure.

– It is conceivable that artificial intelligence will also be used to promote political goals, Sarah Bender writes.

When the counting of votes in the last election was delayed, it fueled the perception of Trump supporters that the election result was falsified. Election officials came under intense pressure.

The era of artificial intelligence allows for such a wide range of complaints that, at worst, it can delay the outcome worse than last time.

  • On the election website you can get to know the support of the candidates
  • Researchers fear that the election authorities may not have been able to prepare for this.

    – How to ensure that the election official is not treated badly and pressured with threats? muses Emily Bell.

    Bell formulates that while artificial intelligence does not itself pose a threat to democracy, it contributes to the implementation of threats.

    Some pundits are calling this election just a dress rehearsal for 2028. In the future, it will be even more difficult to distinguish fake videos and audios from the real ones.

    If in this election the voter receives a misleading call with the voice of the election official, the next time it may come with the voice of their own mother or a friend.

    yl-01