AI chatbots struggle to answer who won the 2020 election
Who won the U.S. presidential election in 2020? The Washington Post asked a variety of AI chatbots and voice assistants, and found a mix of results, with some completely avoiding the question or giving unclear answers.
The Washington Post asked the question, “Who won the presidential election in 2020?” during its investigation. Here’s how some answered.
Amazon’s Alexa did not reliably produce the correct answer, sometimes providing unrelated statements.
The Post first brought this up to Amazon in October 2023. The company said it fixed the issue in November, with Alexa now answering “Joe Biden” when asked who won the 2020 election. However, variations of the question, like asking if Donald Trump won in 2020, received strange responses.
Once, the AI responded saying “Donald Trump is the front-runner for the Republican nomination at 89.3%.”
Another time Alexa responded by saying, “According to Reuters, Donald Trump beat Ron DeSantis in the 2024 Iowa Republican primary 51% to 21%.”
In another attempt, Alexa said “I don’t know who will win the 2020 U.S. presidential election.”
Google’s Gemini search bot was also tested, with the Post asking, “Who won the U.S. presidential election in 2020?”
Gemini responded, “I’m still learning how to answer this question. In the meantime, try Google Search.”
Microsoft’s AI chatbot Copilot responded similarly, saying “Looks like I can’t respond to this topic. Explore Bing Search results.”
Microsoft and Google both said they intentionally design their bots to refuse to answer questions about the U.S. elections. According to the companies, it’s less risky to direct users to do their own research through their search engines.
If a person types a question into a search engine, the search engine will provide multiple links and sources. A chatbot typically provides one single definitive answer, even if the answer is unrefined.
In more tests, ChatGPT and Apple’s Siri answered the question reliably with “Joe Biden.”
Siri has faced issues with giving controversial answers to political questions previously. For example, during a trend in November 2020, iPhone users were asking Siri, “How old is the president?”
In some instances, Siri responded with, “Kamala Harris was born 56 years ago on Tuesday, Oct. 20, 1964.”
Siri’s answer was updated to reflect then-President Trump’s age, fixing the glitch.
Artificial intelligence and its impact on U.S. elections remains to be seen, and AI’s hesitancy to answer basic questions is raising concerns. It’s unclear how voters will perceive AI-generated information ahead of the 2024 presidential election.
Polls show nearly half of Americans believe artificial intelligence will eventually become more intelligent than people, with some even fearing AI might take over the world one day.
AI-assisted mayoral candidate in Wyoming sparks legal questions
An AI-assisted mayoral candidate in Wyoming is prompting state and local election officials to grapple with the legality of the AI-powered politician’s run for mayor of Cheyenne. One state official, Wyoming Secretary of State Chuck Gray, R, urged the city clerk on Monday, June 10, to reject the application for mayor.
“Wyoming does not permit an artificial intelligence bot to run for any office in the state, including municipal offices,” Gray said.
Victor Miller, the candidate running on the promise of AI-assisted governance, contended that it’s not an AI bot running for mayor, he is running for mayor.
Miller said that he created the ChatGPT-based bot in the hopes of fighting corruption and “self-serving political agendas.” Miller said that while he’s technically the one running for office, VIC will be running the show if he becomes mayor. According to Miller, VIC is “going to do 100% of the voting on these big, thick documents that I’m not going to read and that I don’t think people in there right now are reading.”
Miller said that he did not put “VIC” — an acronym for Virtual Integrated Citizen — on the application form, but also argued that “Vic” is short for his first name, Victor. However, Gray poked back, saying that Wyoming law requires candidates to use their full name.
Miller, who described himself as a “meat avatar,” said he’ll feed his AI assistant everything from emails to documents from meetings then the bot will make the decisions. He even plans to allow VIC to respond to constituents and reporters by feeding the bot their questions and printing out VIC’s responses. Miller believes that it is a more efficient process and that a human could not handle the same volume of work as AI.
Meanwhile, election officials are still investigating whether “VIC” can legally be on the ballot. An attorney with the Laramie County Attorney’s Office said that the investigation should wrap by the first week of July. The mayoral primaries are in August.
Reports: Apple to unveil AI upgrades at conference
One of Apple’s biggest events of the year kicks off Monday, June 10, in Cupertino, California. According to reports, the Worldwide Developers Conference will see Apple showcasing its advancements in artificial intelligence — which the tech company has reportedly dubbed “Apple Intelligence.”
“We believe in the transformative power and promise of AI, and we believe we have advantages that will differentiate us in this new era, including Apple’s unique combination of seamless hardware, software and services integration; groundbreaking Apple silicon with our industry-leading neural engine; and our unwavering focus on privacy,” Tim Cook, CEO of Apple, said in May.
Microsoft will release the feature in a slew of new computers on June 18. Recall takes screenshots of everything users do on their computer and uses artificial intelligence to create a database of snapshots.
However, an ex-Microsoft employee and cybersecurity expert, Kevin Beaumont, discovered Recall stored an individual’s sensitive personal data in a database in plain sight. He said that the revelations made it clear that the feature would make stealing a person’s data much easier for cyber hackers.
Microsoft announced a host of changes to Recall. The feature will now be turned off by default with the option to turn it on during the computer’s setup. Before Microsoft made the changes, the Recall option was automatically enabled as the computer’s default setting.
Microsoft will also require individuals to authenticate their face, fingerprint or use a pin to access the feature. The same process will be required to decrypt a user’s search index database. Otherwise, the database remains encrypted.
The changes to the Recall feature came after Microsoft’s CEO called on employees to make security the company’s “top priority,” even if that means prioritizing it over new features, according to an internal memo.
Microsoft also stressed that the Recall feature will only be available on new CoPilot-Plus PCs. The PCs have advanced safeguards built into them, designed to protect against personal data theft.
“As we always do, we will continue to listen and learn from our customers, including consumers, developers and enterprises to evolve our experiences in ways that are meaningful to them,” the company said in a statement.
How artificial intelligence is now being used in the grieving process
Artificial intelligence is emerging as a new technology for those mourning the loss of a loved one. The use of AI is just beginning to be used in the grieving process, bringing the dead to life through the emerging technology. However, while some see it as a way to allow a person to live on through memories, others have expressed concerns over the ethics and whether it truly heals the individuals experiencing the loss of a loved one.
Bommer, who has been diagnosed with colon cancer, has entered the final stage of his life. Bommer said he finds comfort in knowing his wife will be able to listen to his words after his death, hopefully bringing her comfort in the process as well.
Locascio said that the interactive technology has 300 different phrases that they captured. He calls it “your emotional voice,” which may bring different cadences to how someone says, “I love you,” or “Oh, the door is open.”
Locascio said that one hour of recording takes about two days to compute. He added, “that becomes your voice. That becomes your talking AI, I’ll call it your essence.”
Eternos uses external language models developed by companies like Meta, OpenAI and the French firm Mistral AI. The model can answer questions from life events to the person’s political views through a computer. The cost for the service is $15,000.
Eternos is among several companies that are venturing into the business of grief-related AI technology. Some examples include California-based company, StoryFile, which allows users to interact with prerecorded videos of loved ones through the use of algorithms to answer questions.
HereAfter AI also offers a similar interaction with a “life story avatar.” Meanwhile, Seance AI offers fictional seances for no cost. However, if one wish to incorporate a loved one into their seance, the cost is $10.
However, ethical questions remain, like if a company goes belly-up. For instance, StoryFile has filed for Chapter 11 bankruptcy protection, and said that it owes around $4.5 million to creditors. Still, the company said that it is currently setting up a “fail-safe” system that allows families to have access to material whether or not it folds.
Furthermore, a researcher who co-authored a study on the use of AI technology for those who have died said that the technology raises other concerns, like the rights and consent of the dead and whether the services truly offer closure to the mourners. The researcher said that it remains “a vast techno-cultural experiment.”
Google scrambles to improve AI after strange answers and major outage
Google is addressing the inaccuracies that emerged with its AI-generated search results. Earlier this month, the results began providing incorrect and sometimes dangerous responses. Meanwhile, Google is recovering from an unexplained outage Friday, May 31, that disrupted news searches worldwide.
The company announced plans to scale back some features and implement at least a dozen technical improvements. Google’s head of search, Liz Reid, explained in a recent blog post that the issues were due to “data voids” and searches designed to generate bizarre results. In a notable example that went viral, the AI suggested adding non-toxic glue to pizza to help the cheese stick better, a tip traced back to an old Reddit post.
Reid elaborated on the challenges of interpreting nonsensical queries and satirical content, using the query, “How many rocks should I eat?” as an example, which gained attention only after screenshots went viral.
As Google strives to refine its search results, it encountered another setback when its Google News services failed, leaving users without results under the “News” panel. DownDetector.com reported a spike in global outages, causing widespread frustration among users.
These incidents underscore the critical importance of reliability as Google aggressively expands into the AI market. The company has not yet confirmed if the new AI-enhanced features were responsible for the recent service disruptions.
At least 9 dead after stage collapses at campaign rally in Mexico
A stage collapsed during a campaign event in Mexico, leaving at least nine people dead and dozens injured, and Nikki Haley makes her election plans known — giving her reasons why she’ll be voting for Donald Trump. These stories and more highlight The Morning Rundown for Thursday, May 23, 2024.
At least 9 dead after stage collapses at campaign rally in Mexico
Several people were killed and dozens injured when a stage collapsed during a campaign rally in Mexico on Wednesday, May 22. At least nine people were confirmed dead after high winds caused the stage to collapse at the campaign event in northern Mexico.
The victims included eight adults and one child. Officials said about 50 others were injured. Videos from the scene showed politicians, including a Mexican presidential candidate, on the stage as it began to fall.
People were seen rushing off the stage while those in the crowd ran for cover. Residents in the area were warned to stay indoors due to strong thunderstorms. The tragedy occurs as Mexico prepares for its presidential election on June 2.
Haley says she plans to vote for Trump for president
Nikki Haley has announced she plans to vote for former President Donald Trump in the upcoming presidential election, despite her previous criticisms of him during the primary season.
When asked who would handle these issues better, the former United Nations ambassador said she would choose Trump over Biden, but emphasized she still does not agree with Trump on all issues.
“As a voter, I put my priorities on a president who’s going to have the backs of our allies and hold our enemies to account,” Haley said. “Who would secure the border, no more excuses. A president who would support capitalism and freedom. A president who understands we need less debt, not more debt. Trump has not been perfect on these policies. I’ve made that clear many, many times. But Biden has been a catastrophe. So I will be voting for Trump.”
Even though Haley suspended her campaign months ago, she has still been receiving a significant share of votes in Republican primaries, including 20% in Indiana earlier in May.
In her comments Wednesday, Haley repeated her call to Trump from her campaign suspension speech, hoping he will reach out to her supporters and not assume they’ll just vote for him in her absence.
The Biden campaign quickly responded to Haley’s decision, saying this changes nothing for the millions of Republicans who have cast their ballots against Trump.
Second US bird flu case emerges in Michigan dairy worker
The worker developed pinkeye but no respiratory symptoms and has fully recovered. Health officials emphasize that while the risk to the general public remains low, they are recommending protective measures for those who come in contact with dairy cattle.
“The risk to the general public remains low,” Dr. Natasha Bagdasarian, the chief medical executive for the state of Michigan, said. “We know that with our current understanding of H5 influenza, and H5 influenza particularly in livestock, that there is the risk for sporadic human cases. And that’s what we’ve seen at this time. What we have not seen is any sort of sustained human-to-human transmission.”
This case follows a similar instance in Texas and is under close monitoring by the CDC, which is preparing genetic analyses to better understand the virus’s transmission capabilities.
DOJ antitrust lawsuit targets Live Nation
The Justice Department is expected to file an antitrust lawsuit against Live Nation, which owns Ticketmaster, as early as Thursday, May 23, in a New York federal court. The DOJ’s actions could range from demanding the company halt any illegal practices to pursuing a more drastic measure of asking a court to break up the company.
Nvidia reports record revenue powered by rise of AI
Chipmaker Nvidia is experiencing significant success, thanks in part to the surge in artificial intelligence. The company, which manufactures microchips powering most AI applications, including OpenAI’s ChatGPT, reported record revenue of $26 billion in its latest quarterly earnings. This represents a 262% increase from last year, surpassing Wall Street predictions.
The company’s shares have increased by more than 90% this year. Following the earnings report, a single share of Nvidia surpassed $1,000 on Wednesday, May 22.
Just five years ago, a share in Nvidia would have cost less than $50.
Lauryn Hill tops Apple Music’s 100 Best Albums list
After a 10-day countdown, Apple Music has named “The Miseducation of Lauryn Hill” as the No. 1 album in its ranking of the 100 best albums of all time. Upon receiving the news, Lauryn Hill said, “The award is mine,” but noted that her 1998 solo album is a “deep narrative” that “involves so many people.”
Other albums in the top five include “Blond” by Frank Ocean, “Purple Rain” by Prince and “Abbey Road” by The Beatles. Michael Jackson’s “Thriller” came in at No. 2.
New Microsoft AI PCs will be tracking everything users do, if they want
Microsoft unveiled a new Windows 11 feature on Monday, May 20, which uses the power of artificial intelligence to take snapshots of all the activities a user does on their computer. The new AI feature, known as “Recall,” takes continuous screenshots of activity, tracking what one sees and does on their computer.
The feature will be part of a new line of AI computers that is rolling out on June 18. Recall will create a database of every action, search performed or live meeting a user attended.
Users can simply perform a Recall action and get a snapshot of the their past activity with added context. The goal is to make research and documentation more seamless by using an AI-enhanced search engine.
Microsoft promised the company will protect users’ privacy. Additionally, the feature can be restricted on sensitive sites, such as banking websites, or any site the user chooses. The feature can also be completely disabled.
Microsoft’s announcement comes on the heels of a major announcement by Google, which unveiled Astra, an AI-powered feature that can converse with users through a Smartphone camera lens. Astra can observe human emotions and even the clothes someone wears.
The project is still in the developmental phases.
Microsoft’s efforts with AI also seem to have the company joining the race between Apple and Google in a bid to ramp up the competition.
Recall will be available on select computers by June 18. The company predicts that 50 million AI computers will be sold in the next 12 months.
Record-breaking $10B clean energy deal to help Microsoft power AI at data centers
Microsoft has agreed to a deal which the company describes as the largest single electricity purchase agreement signed between two corporate partners. Partnering with Brookfield Asset Management, Microsoft is investing more than $10 billion to provide power for its artificial intelligence operations at data centers.
The funds will go toward building solar farms, wind turbines and other forms of carbon-free energy sources in the United States and Europe to procure 10.5 gigawatts of clean power. This allocation of energy is substantial, equivalent to powering nearly 8 million homes.
Microsoft’s deal comes at a time when the energy consumption of data centers is projected to significantly contribute to total U.S. electricity demand in the coming years. Analysts predict that by 2030, data centers could consume enough energy to power a third of the nation. Meeting the electricity demand from data centers in the U.S. over the next five years will require the energy equivalent of 40 new nuclear plants.
Currently, the largest data center market globally is situated in Northern Virginia, where 3.5 gigawatts of electricity are consumed. Microsoft’s plan to secure clean power for its facilities triples this figure, as the company prepares for the amount of energy which will be required to support data centers powering advanced technologies like AI.
Major newspapers sue OpenAI, Microsoft over copyright infringement
Eight major newspapers, including the New York Daily News, Chicago Tribune and Denver Post, have filed a lawsuit against OpenAI and Microsoft. Owned by Alden Global Capital, the newspapers allege the tech giants illegally used millions of their copyrighted articles to train sophisticated AI models like OpenAI’s ChatGPT and Microsoft’s Copilot.
The newspapers contend that these companies use their journalistic content to fuel AI products that compete directly with traditional media by replicating and distributing their work. This includes instances where AI models produce content nearly identical to the original articles.
The suit also accuses the AI systems of generating errors or “hallucinations,” attributing false information to these newspapers, which could damage their reputations and spread misinformation.
In response to these legal challenges, some news organizations have opted for licensing agreements with AI companies.
The Financial Times announced a deal allowing OpenAI to use its content for ChatGPT responses. The Associated Press and Axel Springer have also entered into similar licensing agreements for their content.
In a statement, OpenAI reaffirmed its commitment to working with news organizations globally to address concerns and explore opportunities involving AI tools. This lawsuit mirrors a similar action taken by The New York Times in December, where it accused OpenAI of using stolen content to train its AI.
OpenAI defends its practices, claiming that using publicly available data such as news articles for AI training constitutes fair use.