Google's Bard will have Limited Answers on Election Queries

Alphabet will reportedly enforce restrictions on types of election queries for its AI chatbot Bard and AI-based searches, along with a slew of preparations ahead the U.S. Presidential elections; the limits are said to apply by early 2024, as per Reuters.

Google did not state what kinds of searches will be blocked. The Big Tech did, however, declare that it will proceed in 2024 with an "abundance of caution" regarding purposes associated with elections, according to Android Central.

Consequently, Android Police adds that Google's Search Generative Experience tool and Bard will also be examined for false information and cybersecurity flaws as they use artificial intelligence to provide context to search results.

Aside from its AI restrictions, the article states that Google's blog post states it will also implement many measures in place to counteract disinformation connected to AI, one of which is a policy requiring advertisers, particularly those involved in political campaigns, to reveal ads that AI has updated.

When adding any AI-generated or edited content to their videos, YouTube producers must also provide this information to viewers. Notably, Google has recently built its own generative AI features, for which it will assume responsibility before the elections.

Google is also reportedly investigating how AI may aid in problem management and how it might propagate false information. The corporation, for example, has regulations to restrict how misinformation, manipulated media, and other factors may undermine democracy. It will now investigate ways to use Large Language Models (LLMs) to support policy enforcement.

The firm also states that political journalists, staff members, campaign workers, and candidates may all enroll in its Advanced Protection Program. The initiative helps protect the most sensitive internet data of prominent people who are often in the public eye. Additionally, Google partnered with Defending Digital Campaigns, a nonpartisan group offering cybersecurity protection.

Big Tech on the Upcoming Elections

Reuters adds that other Big Tech companies have also implemented efforts before the 2024 elections, including Meta, previously stating that it is not allowing marketers in other regulated industries or political campaigns to use its new generative AI advertising solutions.

Additionally, advertisers on Meta will be required to declare any instances in which Facebook and Instagram advertising about politics, social issues, or elections are altered or created using artificial intelligence (AI) or other digital technologies.

Elon Musk's social media site X has also reportedly announced that it will permit political parties and candidates to run ads in the United States. In advance of the American election, it will also grow its elections and safety staff.

AI Risks During the Elections

These preparations for the upcoming elections in various countries follow after Sky News reported experts have warned that the world is unprepared for the impact of AI on the forthcoming elections as it will reportedly be the first elections held in some of the largest democracies in the world since the mainstreaming of generative AI tools such as ChatGPT and picture producers like Midjourney.

The CEO of Project Liberty, a nonprofit organization dedicated to promoting internet safety, Martina Larkin, cautioned that politicians were "at the top of the pyramid" regarding disinformation powered by artificial intelligence.

She adds that in the lead-up to the elections in 2024, this kind of false material may increase much more widely.

Deepfake tapes have also regularly targeted US President Joe Biden, potential rival Donald Trump, and Volodymyr Zelenskyy of Ukraine; also, the voice of Labour leader Sir Keir Starmer was copied for malicious intents.