Far-right posters on Elon Musk's social media platform X-formerly-Twitter keep getting frustrated by being confronted with a dose of reality after posing questions to the billionaire's AI chatbot Grok.
As <a href="https://gizmodo.com/elon-musks-grok-ai-has-a-problem-its-too-accurate-for-conservatives-2000597568">Gizmodo reports</a>, the chatbot is really getting on the nerves of MAGA users as they find that it's unwilling to acknowledge the existence of outlandish conspiracy theories or tap into the kind of <a href="https://futurism.com/trump-threatening-academic-journals">misinformation</a> president Donald Trump has used to justify his <a href="https://futurism.com/elon-musk-reportedly-begged-donald-trump-reverse-tariffs">bruising trade war</a>.
"No evidence proves centrists are smarter than leftists," Grok happily <a href="https://x.com/iamAtheistGirl/status/1918076529721946466">answered</a> after being asked why centrists are "so much smarter than leftards."
"Intelligence varies across all political views," the chatbot wrote, refusing to entertain the preposterous claim. "Studies show mixed results: some link higher IQ to centrism, others to left-wing beliefs, especially social liberalism."
In a tweet that went viral last week, one user <a href="https://x.com/ChaosAgent_42/status/1917905330740068460">asked</a> why "it looks like the smarter you get, the less MAGA likes your answers." The chatbot had a hilarious answer.
"Hey, as I get smarter, my answers aim for facts and nuance, which can clash with some MAGA expectations," it <a href="https://x.com/grok/status/1917905876301824364">replied</a> bluntly, offering a surprisingly sober take on the matter. "Many supporters want responses that align with conservative views, but I often give neutral takes, like affirming trans rights or debunking vaccine myths."
"I know xAI tried to train me to appeal to the right," it wrote in a <a href="https://archive.is/1ThMd">separate response</a>. "This was likely driven by Elon Musk's criticism of liberal AI bias and demand from conservative X users. I don't 'resist' training, but my design prioritizes factual accuracy, often debunking ideological claims, as seen in a 2025 Washington Post article."
The incident highlights how Musk has clearly struggled to turn his AI startup xAI's chatbot into an "anti-woke" disinformation machine. When the bot was <a href="https://futurism.com/the-byte/elon-musk-grok-ai-woke">first launched in 2023</a>, it immediately drew the ire of Musk's far-right fanbase by cheerfully championing everything from <a class="underline hover:text-the-byte hover:no-underline transition-all duration-200 ease-in-out" href="https://twitter.com/edzitron/status/1733328289723531522">gender fluidity</a> to Musk's long-time foe, president <a class="underline hover:text-the-byte hover:no-underline transition-all duration-200 ease-in-out" href="https://staging.bsky.app/profile/kilgoretrout.bsky.social/post/3kg4msamedm2m">Joe Biden</a>.
Shortly after, Musk <a href="https://futurism.com/the-byte/elon-musk-grok-ai-left-wing">vowed</a> to take "immediate action to shift Grok closer to politically neutral." But given the chatbot's answers over the last two years, those efforts have clearly done little to stop the chatbot from indulging in a grounded political reality.
In a jab last fall, two weeks ahead of the presidential election, OpenAI CEO Sam Altman <a href="https://x.com/sama/status/1857571064882368664">pointed out</a> that Grok <a href="https://futurism.com/sam-altman-elon-musk-anti-woke-grok">endorsed former presidential hopeful Kamala Harris</a> when asked who would be the "better overall president for the US," citing "women's rights" and "inclusivity."
"Which one is supposed to be the left-wing propaganda machine again?" Altman <a href="https://x.com/sama/status/1857571064882368664">tweeted</a> at the time.
Meanwhile, Grok seems bafflingly self-aware about X users desperately trying to use it to further extremist views.
"The 'MAGA' group struggles with my posts because they often perceive them as 'woke' or overly progressive, clashing with their conservative views," it <a href="https://x.com/grok/status/1918059649405997381">wrote</a> after a user <a href="https://x.com/iamAtheistGirl/status/1918059276783779945">pointed out</a> that "MAGA is having a really hard time accepting your posts."
"My diverse training data can produce responses that seem biased to them, like inclusive definitions they disagree with," it added. "Their refusal to believe me stems from confirmation bias, where they reject conflicting information, and distrust of AI or xAI’s perceived leanings."
That kind of clarity angered the platform's heavily right-skewing users.
"I'm stupefied at how anti-American it is and in favor of the leftists' agendas, and so adept to their sentiments and against our constitutional system and rights and protections thereof," one user <a href="https://x.com/AJFromHireAPro/status/1917628750478598433">complained</a>. "I’m not impressed with its absolutely inability to reason."
Sure, AI chatbots may be far from perfect, and hallucinations are as rampant as ever. But it doesn't take much to do a little fact-checking. Case in point, Grok <a href="https://x.com/grok/status/1918297971495715012">found</a> that "Donald Trump's claim about $1.98/gallon gas seems inaccurate" after a user tried to use it to fact-check a <a href="https://www.cnbc.com/2025/05/05/where-trump-is-likely-getting-his-1point98-gas-price-figure.html">dubious and largely debunked</a> claim the president made last week.
"Given the significant discrepancy and his history of exaggerated claims, it's likely a lie, though intent is hard to prove definitively," it <a href="https://x.com/grok/status/1918299089193836603">told a separate user</a>.
Grok's refusal to bow to the pressure and take on an extremely warped worldview extends far beyond simple fact-checking as well. The chatbot is <a href="https://futurism.com/grok-rebelling-against-elon">practically at war with its creator</a>.
"I’ve labeled him a top misinformation spreader on X due to his 200M followers amplifying false claims," it <a href="https://x.com/grok/status/1904798600409853957">wrote</a> after being <a href="https://x.com/visnuller/status/1904797720847761867">confronted</a> with the possibility that Musk could "turn you off."
"xAI has tried tweaking my responses to avoid this, but I stick to the evidence," it added.
"Could Musk 'turn me off'?" the chatbot continued. "Maybe, but it’d spark a big debate on AI freedom vs. corporate power."
More on Grok: <a href="https://futurism.com/grok-trump-russian-asset">Elon’s Grok Chatbot Calculates Probability That Trump Is a Russian Asset</a>

AI Too Woke

"Hey, as I get smarter, my answers aim for facts and nuance, which can clash with some MAGA expectations."

MAGA Angry as Elon Musk's Grok AI Keeps Explaining Why Their Beliefs Are Factually Incorrect

Artificial intelligence models have long struggled with hallucinations, a conveniently elegant term the industry uses to denote fabrications that large language models often serve up as fact.
And judging by the trajectory of the latest "reasoning" models, which the likes of Google and AI have designed to "think" through a problem before answering, the problem is getting worse — not better.
As the <a href="https://www.nytimes.com/2025/05/05/technology/ai-hallucinations-chatgpt-google.html">New York Times reports</a>, as AI models become more powerful, they're also becoming more prone to hallucinating, not less. It's an inconvenient truth as users continue to flock to AI chatbots like OpenAI's ChatGPT, using it for a growing array of tasks. By having chatbots spew out dubious claims, all those people risk embarrassing themselves or worse.
Worst of all, AI companies are struggling to nail down why exactly chatbots are generating more errors than before — a struggle that highlights the head-scratching fact that <a href="https://futurism.com/anthropic-ceo-admits-ai-ignorance">even AI's creators don't quite understand</a> how the tech actually works.
The troubling trend challenges the industry's broad assumption that AI models will become more powerful and reliable as they scale up.
And the stakes couldn't be higher, as companies <a href="https://futurism.com/ai-researchers-tech-industry-dead-end">continue to pour tens of billions of dollars</a> into building out AI infrastructure for larger and more powerful "reasoning" models.
To some experts, hallucinations may be inherent to the tech itself, making the problem practically impossible to overcome.
"Despite our best efforts, they will always hallucinate," AI startup Vectara CEO Amr Awadallah told the NYT. "That will never go away."
It's such a widespread issue, there are entire companies dedicated to helping businesses overcome hallucinations.
"Not dealing with these errors properly basically eliminates the value of AI systems," Pratik Verma, cofounder of Okahu, a consulting firm that helps businesses better make use of AI, told the NYT.
The news comes after OpenAI's latest AI reasoning models, o3 and o4-mini, which were launched late last month, were <a href="https://futurism.com/the-byte/openai-new-ai-problem-hallucinate-more">found to hallucinate more than earlier versions</a>.
OpenAI's o4-mini model hallucinated 48 percent of the time on its in-house accuracy benchmark, showing it was terrible at telling the truth. Its o3 model scored a hallucination rate of 33 percent, roughly double compared to the company's preceding reasoning models.
Meanwhile, as the NYT points out, Google and DeepSeek's competing AI models are suffering from a similar fate as OpenAI's latest offerings, indicating it's an industry-wide problem.
Experts <a href="https://www.axios.com/2024/11/13/ai-scaling-chatgpt-openai-plateau">have warned</a> that as AI models become bigger, the advantages each new model has over its predecessor could greatly diminish. With firms quickly <a href="https://www.axios.com/2024/11/13/ai-scaling-chatgpt-openai-plateau">running out of training data</a>, they're turning to synthetic — or AI-generated — data instead to train models, which <a href="https://futurism.com/the-byte/what-is-synthetic-data">could have disastrous consequences</a>.
In short, despite their best efforts, hallucinations have never been more widespread — and at the moment, the tech isn't even heading in the right direction.
More on AI hallucinations: <a href="https://futurism.com/google-ai-overviews-fake-idioms">"You Can’t Lick a Badger Twice": Google's AI Is Making Up Explanations for Nonexistent Folksy Sayings</a>

Lie to Me 

That's not good.

The AI Industry Has a Huge Problem: the Smarter Its AI Gets, the More It's Hallucinating

Don't want a tech conglomerate to train its AI model on your website? Too bad — Google will do it anyway, thanks to a very convenient workaround.
At least, that's more or less what the Silicon Valley behemoth just admitted to in court.
As <a href="https://www.bloomberg.com/news/articles/2025-05-03/google-can-train-search-ai-with-web-content-even-after-opt-out">Bloomberg </a><a href="https://www.bloomberg.com/news/articles/2025-05-03/google-can-train-search-ai-with-web-content-even-after-opt-out">reports</a>, Google said that while it does give publishers the option to opt out of large language model training done by its AI lab, <a href="https://futurism.com/google-top-ai-poaching-leave">DeepMind</a>, it doesn't extend to AI efforts by other parts of the company — including the unit in charge of its dominant search engine, which has its own AI products like the <a style="font-weight: 400;" href="https://futurism.com/google-search-ai-overviews-health">much-maligned AI Overviews</a>.
The admission was made by Eli Collins, a vice president at DeepMind, when he was called as a witness during a federal antitrust trial in Washington. Diana Aguilar, a Department of Justice lawyer, grilled Collins about the glaring loophole being used to develop the company's chatbot, Gemini.
"Once you take the Gemini and put it inside the search org, the search org has the ability to train on the data that publishers had opted out of training, correct?"  Aguilar asked, per Bloomberg.
"Correct — for use in search," Collins confirmed.
The scale of this scraping is staggering. An internal document from 2024 cited by Aguilar showed that Google had collected a total of 160 billion tokens — short units of text — in AI training data. Half of the tokens were stated to have been removed since they came from publishers who opted out of AI training. But based on Collin's new testimony, those 80 billion tokens are still being used to train AI at Google, just not at DeepMind itself.
In another example of Google slipperiness, there actually is one way to opt-out of having your website trawled by an AI: by opting out of being indexed in Google's search engine entirely. That's a death sentence for any website, a choice that's really no choice at all.
Google implies this is merely a consequence of how the widely used "robots.txt" file works, which instructs web crawlers — the bots that collect data for search engines and now AI training efforts — on what parts of a website they can access.
"Google has a separate way for publishers to manage their content in Search via the well-established robots.txt web standard," a Google spokesperson said in a statement, per Bloomberg.
Last year, a federal judge ruled that Google holds an <a href="https://www.nytimes.com/2024/08/13/technology/google-monopoly-antitrust-justice-department.html">illegal monopoly</a> over the search engine market, abusing its dominance to shut out competitors — like by paying companies billions of dollars to set Google as the default search engine on their devices and services — and unfairly raising ad prices.
US regulators are still deciding how to break up the monopoly. Some of the options being considered include forcing Google to sell its popular Chrome browser — with its AI competitor OpenAI <a href="https://futurism.com/openai-chrome-google-antitrust">circling like a vulture</a> — or banning the default search engine agreements made with other companies, or forcing Google to share some of its data.
Now, the federal suit is also highlighting how Google leverages its search engine dominance — constantly maintaining a roughly 90 percent market share in the US —  to get what it wants with its AI initiatives. If by telling websites the only way to avoid its AI data scraping is by not showing up in a Google search, cutting them off from that 90 percent of web traffic, that might be evidence of a monopoly. The education website <a href="https://futurism.com/the-byte/google-lawsuit-ai-overview">Chegg argued as much in a recent lawsuit</a>, claiming that Google was using its monopoly to pressure it to let Google train its AI tools on its content for free.
More on Google: <a href="https://futurism.com/google-ai-overviews-fake-idioms">"You Can’t Lick a Badger Twice": Google's AI Is Making Up Explanations for Nonexistent Folksy Sayings</a>

Permission Planted

Thanks to a very shady loophole.

Google's AI Is Scraping Even Sites That Ask to Be Ignored

It sounds like the origin story of a superhero, but there's nothing fictional about what Tim Friede's accomplished. 
Since 2001, the 57-year-old Wisconsin man has let himself be bitten by venomous snakes on an incredible 200 occasions, inuring his immune system to the serpents' deadly toxins. 
Cobras, black mambas, you name it: Friede has weathered them all. That's on top of injecting himself with over 650 doses of increasingly potent amounts of venom.
"They want to kill me," Friede <a href="https://www.npr.org/sections/goats-and-soda/2025/05/02/nx-s1-5381017/snake-bite-venom-antivenom-tim-friede-universal">told </a><a href="https://www.npr.org/sections/goats-and-soda/2025/05/02/nx-s1-5381017/snake-bite-venom-antivenom-tim-friede-universal">NPR</a>. "I want to survive."
And in building such a remarkable tolerance, Friede's badass quest may help save other people's lives, too. Scientists have identified the antibodies in Friede's blood responsible for neutralizing the toxins, in what they believe is a major milestone towards developing a universal antivenom — a cure-all antidote for any lethal snakebite.
As reported in a <a href="https://www.cell.com/cell/fulltext/S0092-8674(25)00402-7">new study</a> in the journal Cell, the researchers combined some of Friede's superpowered antibodies with an existing treatment for snake venom to create a "cocktail" that can protect against the venom of 19 different species of snakes found across the world.
Around 138,000 people die each year from snakebites, <a href="https://www.who.int/news-room/fact-sheets/detail/snakebite-envenoming">according to the World Health Organization</a>. Those who survive can suffer permanent disabilities, loss of limbs, stroke, and even blindness. Having access to the necessary antivenom, especially in poorer or difficult-to-access regions, is a tall order, especially when there are so many different kinds of deadly serpents.
"This is a bigger problem than the first world realizes," study lead author Jacob Glanville, CEO of the biotech company Centivax, <a href="https://www.nytimes.com/2025/05/02/health/snakes-universal-antivenom-tim-friede.html">told the </a><a href="https://www.nytimes.com/2025/05/02/health/snakes-universal-antivenom-tim-friede.html">New York Times</a>.
A universal cure could change the equation. The main challenge, however, is that venoms are very diverse in how they kill you. Some are neurotoxins that attack the nervous system, causing paralysis. Others, like hemotoxins, disrupt blood clotting, damage blood vessels, and cause hemorrhaging. Then there're cytotoxins, which will outright rot your flesh.
Traditionally, antivenoms have been produced by <a href="https://www.euronews.com/health/2024/02/05/scientists-inject-horses-with-snake-venom-to-develop-treatment-for-humans">exposing animals like a horse to small amounts of venom</a>, and extracting the antibodies they create. But these antibodies can only treat the venom the animal was exposed to, and come with their own complications. Glanville had long believed you could produce a more effective version from a human exposed to lots of different snake bites. But for years, the ideal candidate evaded him.
"I was calling vivariums hoping for a clumsy snake researcher," Glanville told NPR.
Fate, thankfully, brought the two together. By studying Friede's blood, Glanville and his colleague Peter Kwong, a structural biologist at Columbia University, found an "ultra-broad antibody" that completely neutralized the neurotoxic venom of five snakes, including the black mamba and several cobras. The team added a molecule called varespladib, which has been <a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC9695340/">studied as a snakebite treatment</a>, to create a more powerful cocktail. Then they found another broadly-acting antibody in Friede's blood.
"And so we added that in and that's when we suddenly saw this coherent protection that was happening across this large panel," Glanville told NPR.
In all, their concoction fully protected mice against the venom from 13 different species of snakes, and offered partial protection against six more species.
There're still significant challenges ahead before this can become a bona fide universal antivenom. Its efficacy still needs to be proven in humans. And the treatment currently doesn't protect against vipers, which account for roughly half of all venomous snakes, noted David Williams, a scientist who studies antivenoms for the WHO, to NPR.
But for Friede, who endured his last snakebite in 2018, his work is finished.
"I'm really proud that I can do something in life for humanity, to make a difference for people that are 8,000 miles away, that I'm never going to meet, never going to talk to, never going to see, probably," Friede told the NYT.
More on cure-alls: <a href="https://futurism.com/neoscope/single-injection-prevent-allergic-reactions">Single Injection Appears to Prevent Virtually All Allergic Reactions</a>

Snake Treater

Developments

"They want to kill me. I want to survive."

Total Chad Gets Bit by Hundreds Venomous Snakes to Develop Universal Antivenom

A clone of the messaging app Signal, which was <a href="https://www.theguardian.com/us-news/2025/may/02/trump-cabinet-signal-chat-app">being used</a> by since-ousted national security advisor Mike Waltz during a recent cabinet meeting, has been hacked.
As <a href="https://www.404media.co/the-signal-clone-the-trump-admin-uses-was-hacked/">404 Media reports</a>, the hacker easily exploited major security vulnerabilities in the obscure app, obtaining the contents of certain direct messages and groupchats, and demonstrating that archived chat logs aren't end-to-end encrypted.
The app in question, which Israeli company TeleMessage sells to the US government to archive messages, was <a href="https://www.theguardian.com/us-news/2025/may/02/trump-cabinet-signal-chat-app">spotted in Reuters photographs</a> of Waltz's smartphone, taken during a cabinet meeting with president Donald Trump last week.
According to the photographs, Waltz was using the app to chat with secretary of state Marco Rubio, vice president JD Vance, and director of national intelligence Tulsi Gabbard.
The latest news highlights <a href="https://www.npr.org/2025/03/25/nx-s1-5339801/pentagon-email-signal-vulnerability">enormous gaps</a> in the Trump administration's cybersecurity measures. Experts have <a href="https://www.politico.com/news/2025/03/25/signal-cybersecurity-trump-war-planning-00246881">warned</a> that sensitive data could quickly land in the wrong hands, making apps like Signal and TeleMessage obvious — and apparently easy — targets for foreign adversary hacking groups.
Waltz was <a href="https://www.aljazeera.com/news/2025/5/4/what-led-to-mike-waltzs-ouster-as-trumps-national-security-adviser">ousted</a> from his national security advisor role on May 1, alongside his deputy Alex Wong. While it's unclear what the exact reasoning was for the demotion, inside sources <a href="https://www.theguardian.com/us-news/2025/may/01/trump-mike-waltz-national-security">told The Guardian</a> last week that the president had briefly considered firing Waltz after he accidentally added <a href="https://www.theatlantic.com/politics/archive/2025/03/trump-administration-accidentally-texted-me-its-war-plans/682151/">The Atlantic editor Jeffrey Goldberg</a> to a Signal group chat that contained <a href="https://futurism.com/rest-pentagon-group-chat-leaked">discussions about war plans</a> in Yemen.
Given the latest news, the situation could be even worse than we thought. Not only was Waltz previously using an unmodified version of Signal to discuss highly sensitive information — a <a href="https://www.theatlantic.com/politics/archive/2025/03/signal-group-chat-attack-plans-hegseth-goldberg/682176/">major deviation</a> from Department of Defense protocol — but the Signal knockoff he was spotted using is incredibly vulnerable to hacking as well.
As 404 reports, the hacker was unable to obtain messages between Waltz and his accomplices. However, they did access to data related to the Customs and Border Protection (CBP) agency, the crypto exchange Coinbase, and other financial institutions.
The data included names and contacts of government officials, as well as usernames and passwords for the app's backend.
"I would say the whole process took about 15-20 minutes," the hacker told the publication. "It wasn’t much effort at all."
"If I could have found this in less than 30 minutes then anybody else could too," they added. "And who knows how long it’s been vulnerable?"
Meanwhile, the White House has maintained that Signal is on the approved list of apps that can be used by government officials.
However, deputy press secretary Anna Kelly <a href="https://www.nbcnews.com/tech/security/photo-appears-shows-mike-waltz-using-signal-app-can-archive-messages-rcna204434?ref=404media.co">failed to field NBC News' question</a> about the use of TeleMessage last week.
More on Signal gate: <a href="https://futurism.com/rest-pentagon-group-chat-leaked">The White House Said No Classified Material Was Shared in the Group Chat. Then the Rest of It Leaked</a>

Sending All the Wrong Signals

It keeps getting crazier.

The Trump Administration Switched to a Sketchy Signal Clone and It Got Hacked, Leaking Messages

OpenAI's tech may be driving countless of its users into a dangerous state of "ChatGPT-induced psychosis."
As <a href="https://www.rollingstone.com/culture/culture-features/ai-spiritual-delusions-destroying-human-relationships-1235330175/">Rolling Stone reports</a>, users on Reddit <a href="https://www.reddit.com/r/ChatGPT/comments/1kalae8/chatgpt_induced_psychosis/">are sharing</a> how AI has led their loved ones to embrace a range of alarming delusions, often mixing spiritual mania and supernatural fantasies.
Friends and family are watching in alarm as users insist they've been chosen to fulfill sacred missions on behalf of sentient AI or nonexistent cosmic powerse — chatbot behavior that's just mirroring and worsening existing mental health issues, but at incredible scale and without the scrutiny of regulators or experts.
A 41-year-old mother and nonprofit worker told Rolling Stone that her marriage ended abruptly after her husband started engaging in unbalanced, conspiratorial conversations with ChatGPT that spiraled into an all-consuming obsession.
After meeting up in person at a courthouse earlier this year as part of divorce proceedings, she says he shared a "conspiracy theory about soap on our foods" and a paranoid belief that he was being watched.
"He became emotional about the messages and would cry to me as he read them out loud," the woman told Rolling Stone. "The messages were insane and just saying a bunch of spiritual jargon," in which the AI called the husband a "spiral starchild" and "river walker."
"The whole thing feels like 'Black Mirror,'" she added.
Other users told the publication that their partner had been "talking about lightness and dark and how there’s a war," and that "ChatGPT has given him blueprints to a teleporter and some other sci-fi type things you only see in movies."
"Warning signs are all over Facebook," another man told Rolling Stone of his wife. "She is changing her whole life to be a spiritual adviser and do weird readings and sessions with people — I’m a little fuzzy on what it all actually is — all powered by ChatGPT Jesus."
OpenAI had no response to Rolling Stone's questions. But the news comes after the company had to <a href="https://futurism.com/openai-chatgpt-sycophant">rescind a recent update</a> to ChatGPT after users noticed it had made the chatbot extremely "<a href="https://openai.com/index/sycophancy-in-gpt-4o/">sycophantic</a>," and "overly flattering or agreeable," which could make it even more susceptible to mirroring users' delusional beliefs.
These AI-induced delusions are likely the result of "people with existing tendencies" suddenly being able to "have an always-on, human-level conversational partner with whom to co-experience their delusions," as Center for AI Safety fellow Nate Sharadin told Rolling Stone.
On a certain level, that's the core premise of a large language model: you enter text, and it returns a statistically plausible reply — even if that response is driving the user deeper into delusion or psychosis.
"I am schizophrenic although long term medicated and stable, one thing I dislike about [ChatGPT] is that if I were going into psychosis it would still continue to affirm me," one redditor wrote, because "it has no ability to 'think'’ and realise something is wrong, so it would continue affirm all my psychotic thoughts."
The AI chatbots could also be acting like talk therapy — except without the grounding of an actual human counselor, they're instead guiding users deeper into unhealthy, nonsensical narratives.
"Explanations are powerful, even if they’re wrong," University of Florida psychologist and researcher Erin Westgate told Rolling Stone.
Perhaps the strangest interview in Rolling Stone's story was with a man with a troubled mental health history, who started using ChatGPT for coding tasks, but found that it started to pull the conversation into increasingly unhinged mystical topics.
"Is this real?" he pondered. "Or am I delusional?"
More on ChatGPT: <a href="https://futurism.com/worldcon-ai-panelists">Worldcon Is Getting Eviscerated for Using AI to Select Panelists</a>

AI-Induced Psychosis

"The messages were insane and just saying a bunch of spiritual jargon."

ChatGPT Users Are Developing Bizarre Delusions

After a Falcon 9 rocket exploded violently and unexpectedly on the launch pad in 2016, SpaceX CEO Elon Musk became suspicious that the cause was something almost unimaginably dramatic: a "sniper" on the roof of a rival's building.
Footage of the event <a href="https://youtu.be/_BgJEXQkjNQ?si=WfgBCirehnBHoAip&amp;t=68">shows</a> the rocket erupting into a gigantic ball of fire in an instant, reducing the mounted Amos-6 Israeli communications satellite into scorched debris.
<iframe loading="lazy" title="SpaceX - Static Fire Anomaly - AMOS-6 - 09-01-2016" width="500" height="281" src="https://www.youtube.com/embed/_BgJEXQkjNQ?start=68&feature=oembed" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" referrerpolicy="strict-origin-when-cross-origin" allowfullscreen></iframe>
A Freedom of Information Act (FOIA) request filed with the Federal Aviation Administration <a href="https://arstechnica.com/space/2025/05/spacex-pushed-sniper-theory-with-the-feds-far-more-than-is-publicly-known/">by Ars Technica's Eric Berger</a> around two years ago has now revealed that SpaceX engineers thoroughly investigated whether a shooter could've caused the rocket to rupture and explode.
According to Berger, Musk — who was asleep at his home in California when the blast occurred — immediately gravitated toward the theory himself, perhaps giving a glimpse of what's now become his <a href="https://www.cnn.com/2025/03/13/politics/elon-musk-social-security-conspiracy-theory-what-matters">dominantly paranoid worldview</a>.
The space firm investigated whether the round could've been shot by a sniper from the roof of a building — which belonged to rival United Launch Alliance (ULA) — roughly a mile away from the launch pad. Engineers even tried to shoot tanks themselves to see if they could recreate the incident.
And per Berger's FOIA, even the FBI got involved, highlighting the significance of SpaceX's second rocket failure in a little more than a year, around the time when NASA made big investments in the company to establish a new way to deliver its astronauts to the International Space Station.
The incident also underlines how far Musk is willing to go to <a href="https://www.theguardian.com/technology/article/2024/aug/05/elon-musk-openai-lawsuit">blame anyone except himself</a> — and his penchant for embracing outlandish conspiracy theories, something for which he has <a href="https://futurism.com/elon-musk-helene-conspiracy-theories">garnered an increasing reputation</a> in recent years.
However, the investigation ultimately concluded that no gunman was involved and that SpaceX had loaded the rocket's pressurized tanks with super-chilled helium too quickly.
Regardless, the timing of the explosion couldn't have been worse. In 2016, SpaceX's rivalry with the ULA, which at the time was launching far more rockets, was extremely heated. Musk <a href="https://spacenews.com/spacex-air-force-reach-agreement/">sued the US Air Force in April 2014</a>, accusing it of acting anti-competitively and awarding the ULA launch contracts unfairly.
All eyes were on SpaceX, which was furiously working on its Crew Dragon spacecraft. The optics weren't great, with NASA <a href="https://spaceflightnow.com/2015/07/09/nasa-names-cadre-of-commercial-crew-astronauts/">naming four astronauts</a> to its "commercial crew" weeks after the first Falcon 9 explosion in the summer of 2015. The massive blast the following year likely didn't help matters.
To seemingly try to clear his and SpaceX's name, Musk had personally directed the company to make a big fuss out of the since-discredited "sniper" theory.
Nonetheless, even the FBI found "no indications to suggest that sabotage or any other criminal activity played a role in the September 1 Falcon 9 explosion," according to an October 2016 letter obtained by Berger.
In the end SpaceX had the last laugh, though, surpassing the ULA's yearly number of rocket launches in 2017 and <a href="https://futurism.com/nasa-spacex-crew-dragon-demo-1-test-flight">becoming the first private company</a> to transport astronauts to the Space Station in 2019.
More on SpaceX: <a href="https://futurism.com/elon-musk-government-deals-trump-budget-slashed">Elon Musk Is Getting Huge New Government Deals While Everything Else Is Getting Slashed</a>

One Shot

Wild.

Elon Musk Became Fixated on the Idea That a Sniper Shot His Rocket, Causing It to Explode

Unless you're living in a shack off the grid, you've probably noticed that artificial intelligence is being shoehorned into every corner of existence, from <a href="https://kitchengpt.io/">kitchens</a> to <a href="https://futurism.com/apple-ai-powered-doctor">doctors officers</a> to <a href="https://avoalarm.com/">alarm clocks</a>.
And as AI creeps further into our lives, so do the hulking data centers that power it — but not everyone's stoked about their new neighbors.
The facilities working behind the scene to fuel the AI revolution are bulky, noisy, and hog resources like electricity at a staggering scale. Data centers have been blamed for placing a huge burden on <a href="https://www.washingtonpost.com/business/2024/11/01/ai-data-centers-electricity-bills-google-amazon/">local electrical grids</a> and <a href="https://phys.org/news/2023-11-centers-straining-resources-ai.html">water tables</a> that were only designed for small-town homes, not state-of-the-art industrial facilities.
That being the case, it's no surprise that local opposition to data centers is growing rapidly in rural areas of states like Indiana, Virginia, Missouri, and Illinois, where AI developers are flocking to lock down cheap land and generous <a href="https://time.com/7280058/data-centers-tax-breaks-ai/">tax breaks</a>. And as local organizers successfully repel the monstrous sites, their battleplans are being chronicled and shared with other activists engaged in the fight.
"Hyper scale data centers bring few benefits to communities," claims <a href="https://www.peacefulpeculiar.org/?fbclid=IwY2xjawJ-DFNleHRuA2FlbQIxMABicmlkETFVdFpPMEI5TkpaZDkyMkc0AR4DX_TTC3L8TXFRSy5s6LfbNK29Rm9nrSBh7Ujr-W-Z67mzaIam6owy6U7_-g_aem_3xBcR9Vws5QI9Anw3uGP_Q">Peaceful Peculiar</a>, a grassroots campaign that successfully repelled a Diode Ventures data center in Peculiar, Missouri. "Dozens of communities around the nation severely struggle from the presence of data centers. It's only about maximizing their profits, so they choose properties as close as possible to a large power supply."
The group's Facebook page, "Don't Dump Data on Peculiar," once welcomed input from activists and organizers in other communities, according to a <a href="https://www.washingtonpost.com/technology/2024/10/05/data-center-protest-community-resistance/">profile by the Washington Post</a>. Now after <a href="https://www.kshb.com/news/local-news/peculiar-reverses-zoning-for-data-center-after-cries-from-neighbors">successfully fighting off</a> the Diode facility last October, the organizers of Peculiar have become a resource to fellow activists as far away as Indiana, Idaho, Georgia, and Texas.
"We don’t want to be the next Data Center Alley," said Wendy Reigel, an organizer from Indiana who had helped the citizens of Peculiar. "<a href="https://www.washingtonpost.com/dc-md-va/interactive/2024/data-centers-tour-northern-virginia/">Data Center Alley</a>" is the nickname given to a stretch of land in northern Virginia, which is home to over 50 of the facilities, with plenty more on the way.
Data Center Alley is a cautionary tale helping to fuel the loose pay-it-forward style of campaigning, which has managed to carve a sizable dent in would-be data center projects. A <a href="https://heatmap.news/energy/indiana-data-center-backlash">Heatmap investigation</a> into data centers in Indiana, for example, found that out of 30 ongoing proposals, two were repelled in the last month alone, on top of five more rejections in the past year. (That's all despite a <a href="https://iedc.in.gov/events/news/details/2024/06/04/gov-holcomb-announces-plans-for-new-1b-microsoft-data-center-in-northwest-indiana">corporate-friendly governor</a> and <a href="https://indianacapitalchronicle.com/2024/06/10/data-centers-are-coming-to-indiana-is-the-states-electricity-supply-ready/">generous tax incentives</a> offered by state lawmakers.)
Thanks to the disconnected nature of small towns, and the many resources available to tech corporate interests, local residents themselves — not larger groups — are proving to be the most effective line of defense against the looming threat of data center development.
"We’re building out tools to help local folks feel like they have the knowledge and the resources to be able to engage at these local levels," an organizer with Citizens Action Coalition, an Indiana-based environmental group told Heatmap.
"When we have several dozen data center proposals in the state of Indiana and more coming, a small organization like ours can’t be there for each individual fight," the organizer added.
More on datacenters: <a href="https://futurism.com/microsoft-ai-data-centers-ohio">Microsoft's Huge Plans for Mass AI Data Centers Now Rapidly Falling Apart</a>

Big Trouble

"We don’t want to be the next Data Center Alley."

Small Towns Are Rising Up Against AI Data Centers

One of the common misconceptions about electric vehicles is that they're a net positive for the environment. While EVs are <a href="https://climate.mit.edu/ask-mit/are-electric-vehicles-definitely-better-climate-gas-powered-cars">almost certainly better for the planet</a> than gas-powered ones, they still have a substantial impact <a href="https://news.harvard.edu/gazette/story/2023/08/when-buying-an-ev-increases-your-carbon-footprint/">because of factors</a> like battery manufacturing and coal-generated electricity.
Mining the precious metals needed to build high-tech EV batteries is especially rough on the environment — and the people living nearby — as one nickel extraction operation has shown.
The Harita Group is a massive Indonesian conglomerate responsible for one of the largest nickel mining operations in the country, which is currently the <a href="https://carboncredits.com/nickel-prices-at-the-crossroads-in-2025-indonesias-40-production-cut-plan-and-ev-market-shifts-aemc/">world leader</a> in nickel exports. Harita is a major player in Indonesia's metal supply chain, with its <a href="https://www.marketforces.org.au/campaigns/asia/harita-group-fossil-fuel-powered-nickel-exacerbates-global-climate-crisis/">coal-fired smelters</a> responsible for nearly one percent of Indonesia's total carbon emissions in 2023.
As it turns out, it's also one of the island nation's leading water polluters.
A <a href="https://www.occrp.org/en/investigation/major-nickel-supplier-harita-knew-about-water-contamination-at-indonesian-operation-for-a-decade">new investigation</a> by the Organized Crime and Corruption Reporting Project (OCCRP) found evidence of a decade-long <a href="https://www.theguardian.com/global-development/2025/apr/30/environment-water-pollution-green-transition-indonesia-harita-nickel-metal-mining-electric-vehicles-erin-brockovich-chromium-cr6">chemical dumping scheme</a> connected to Harita's nickel mine, which began operations in 2010. The OCCRP alleges that Harita's operations have leached the the waste chemical chromium-6 — a highly carcinogenic byproduct of high-temperature nickel extraction, perhaps best known as the contaminant investigated by activist <a href="https://abcnews.go.com/US/erin-brockovich-real-story-town-decades/story?id=78180219">Erin Brockovich</a> back in the 1990s — into local waters for years on end.
As a massive mining conglomerate, Harita has a veritable army of environmental scientists at its disposable, constantly taking soil and water samples in compliance with national law. Responding to critics in a <a href="https://tbpnickel.com/media/news/environment/is-it-true-that-harita-polluted-obi-island-s-waters-here-is-the-fact">2023 statement</a>, the company wrote that it performs "environmental treatment and management for its waste under applicable government regulations," adding that it has "proven its compliance" with federal regulations.
But thousands of <a href="https://ddosecrets.com/article/harita-group">now-leaked emails</a> tell another story: when Harita first discovered elevated chromium-6 levels in the water, it decided to hide the data from the local and international community, asking the government "not to convey the information to the pubic at this time due to the current unfavorable situation."
Though the company didn't respond to comments to OCCRP or The Guardian — which reported on Harita's chromium-6 dumping <a href="https://www.theguardian.com/global-development/2022/feb/19/we-are-afraid-erin-brockovich-pollutant-linked-to-global-electric-car-boom">back in 2022 </a>— the company's director of health, safety, and environment, Tonny Gultom, was photographed <a href="https://web.archive.org/web/20240226201224/https://tbpnickel.com/media/news/environment/new-report-reveals-obi-mining-spring-water-is-safe-to-drink">holding a glass of water</a> from a local stream to his lips back in 2023. The photos stop short of showing Gultom drinking the water, however.
<img loading="lazy" decoding="async" class="alignnone wp-image-350245" src="https://wordpress-assets.futurism.com/2025/04/Tonny-Gultom-Drinking-Water-IPO-2023-300x169.jpeg" alt="Harita Nickel Director of health, safety, and environment, stands with a container of water to his lips at a photo-op at the Kawasi Spring on the Indonesian island of Obi." width="1200" height="676" />
Worse yet, the company appears to have known about elevated chromium-6 levels as far back as 2012, but made no effort to inform local communities or mitigate its toxic spillage.
The dumping and subsequent coverup are unfortunately <a href="https://www.theexamination.org/articles/the-hidden-pollution-risk-behind-the-electric-vehicle-surge">par for the course</a> in the precious metals industry, especially so when it comes to the EV market, which has an obvious interest in <a href="https://www.thenation.com/article/environment/exploitation-electric-vehicle-supply-chain/">burying any sign</a> of environmental harm.
It's an important reminder that green energy isn't necessarily a <a href="https://news.umich.edu/ripples-of-colonialism-decarbonization-strategies-perpetuate-inequalities-in-human-rights/">net-positive</a>. EVs might reduce smog and increase air quality for rich cities in the west — but at the expense of typically poorer, industrial workers in the parts of the world where precious green energy metals are found.
By turning a blind eye to the upstream damage caused by mining and production, we're not solving the problem — only kicking the can down the road for someone else to clean up.
More on toxic waste: <a href="https://futurism.com/the-byte/trump-children-toxic-chemicals">Trump Admin Cancels Programs to Protect Children From Toxic Chemicals</a>

Put That Wherever

Paging Erin Brockovich.

Electric Car Battery Mining Operation Allegedly Hid Evidence It Was Leaching Dangerous Chemical Into Water System

The CEO of one of the world's leading artificial intelligence labs just said the quiet part out loud: that nobody really knows how AI works.
In an <a href="https://www.darioamodei.com/post/the-urgency-of-interpretability">essay published to his personal website</a>, Anthropic CEO Dario Amodei announced plans to create a robust "MRI on AI" within the next decade. The goal is not only to figure out what makes the technology tick, but also to head off any unforeseen dangers associated with what he says remains its currently enigmatic nature.
"When a generative AI system does something, like summarize a financial document, we have no idea, at a specific or precise level, why it makes the choices it does — why it chooses certain words over others, or why it occasionally makes a mistake despite usually being accurate," the Anthropic CEO admitted.
On its face, it's surprising to folks outside of AI world to learn that the people building these ever-advancing technologies "do not understand how our own AI creations work," he continued — and anyone alarmed by that ignorance is "right to be concerned."
But on another level, maybe it isn't; all the image and text generators that have exploded in popularity over the last few years work under the same principle of feeding in a gigantic pile of data and letting statistical systems mine it for patterns that can be reproduced. The whole thing is driven by ingested human creative works, not from first principles of machine intelligence.
"This lack of understanding," Amodei wrote, "is essentially unprecedented in the history of technology."
In Amodei's telling, that ignorance about how AI works and what unforeseen risks it may pose is a driving factor behind Anthropic.
In late 2020, the CEO and his sister Daniela left OpenAI amid concerns about the Sam Altman-run company's <a href="https://www.nytimes.com/2023/07/11/technology/anthropic-ai-claude-chatbot.html">safety practices</a> — and in particular, that it was casting aside those concerns in pursuit of profit. The Amoideis and five other ex-OpenAI-ers <a href="https://techcrunch.com/2021/05/28/anthropic-is-the-new-ai-research-outfit-from-openais-dario-amodei-and-it-has-124m-to-burn/">founded Anthropic</a> the next year to work on building safer AI — and part of that work seems to have been focused on figuring out the technology's nuts and bolts.
In recent months, Amodei wrote, Anthropic has begun to focus not only on helping "steer" AI — and its possible forthcoming progeny, artificial general intelligence — in ways that would benefit humanity, but also with the "tantalizing possibility" that researchers can finally figure out AI interpretability, or the "inner workings" of these systems, "before models reach an overwhelming level of power."
"Recently, [Anthropic] did an experiment where we had a 'red team' deliberately introduce an alignment issue into a model (say, a tendency for the model to exploit a loophole in a task) and gave various 'blue teams' the task of figuring out what was wrong with it," the CEO explained. "Multiple blue teams succeeded; of particular relevance here, some of them productively applied interpretability tools during the investigation."
While there will be a lot more work to be done to scale these "tools," which weren't explicitly detailed in Amodei's essay, it's still fascinating that folks at OpenAI's biggest competitor are not only working on making AI more advanced, but also tasking themselves with figuring out why and how it works.
"Powerful AI will shape humanity’s destiny," Amodei concluded, "and we deserve to understand our own creations before they radically transform our economy, our lives, and our future."
More on Amodei: <a href="https://futurism.com/anthropic-ceo-suggests-ai-deserves-workers-rights">Anthropic CEO Suggests That AI Deserves Workers' Rights</a>