Elon Musk, CEO of Tesla, speaks with CNBC on May 16th, 2023.
David A. Grogan | CNBC
The sudden departure of Twitter executives tasked with content moderation and brand safety has left the company more vulnerable than ever to hate speech.
On Thursday, Twitter’s vice president of trust and safety, Ella Irwin, resigned from the company. Following Irwin’s departure, the company’s head of brand safety and ad quality, A.J. Brown, reportedlyleft, as did Maie Aiyed, a program manager who worked on brand safety partnerships.
It’s been just over seven months since Elon Musk closed his $44 billion purchase of Twitter, an investment that’s so far has been a giant money loser. Musk has dramatically downsized the company’s workforce and rolled back policies that restricted what kind of content could circulate. In response, numerous brands suspended or decreased their advertising spending, as several civil rights groups have documented.
Twitter, under Musk, is the fourth most-hated brand in the U.S. according to the 2023 Axios Harris reputation rankings.
The controversy surrounding Musk’s control of Twitter continues to build.
This week, Musk said it’s not against Twitter’s terms of service to misgender trans people on the platform. He said doing so is merely “rude” but not illegal.” LGBTQ+ advocates and researchers dispute his position, claiming it invites bullying of trans people. On Friday, Musk boosted a video on Twitter that was deemed transphobic by these groups.
Numerous LGBTQ organizations expressed dismay to NBC News over Musk’s decision, saying the company’s new policies will lead to an uptick in anti-trans hate speech and online abuse.
Although Musk recently hired former NBC Universal global advertising chief Linda Yaccarino to succeed him as CEO, it’s unclear how the new boss will assuage advertisers’ concerns regarding racist, anti-Semitic, transphobic and homophobic content in light of the recent departures and Musk’s ongoing role as majority owner and technology chief.
Even before the latest high-profile exits, Musk had been reducing the number of workers tasked with safety and content moderation as part of the company’s widespread layoffs. He eliminated the entire AI ethics team, which was responsible for ensuring that harmful content wasn’t being algorithmically recommended to users.
Musk, who is also the CEO of Tesla and SpaceX, has recently downplayed concerns about the prevalence of hate speech on Twitter. He claimed during a Wall Street Journal event that since he took over the company in October, hate speech on the platform has declined, and that Twitter has slashed “spam, scams and bots” by “at least 90%.”
Experts and ad industry insiders told CNBC that there’s no evidence to support those claims. Some say Twitter is actively impeding independent researchers who are attempting to track such metrics.
Twitter didn’t provide a comment for this story.
The state of hate speech on Twitter
In a paper published in April that will be presented at the upcoming International Conference on Web and Social Media in Cyprus, researchers from Oregon State, University of Southern California, and other institutions showed that hate speech has increased since Musk bought Twitter.
The authors wrote that the accounts known for posts containing hateful content and slurs targeting Blacks, Asians, LGTBQ groups and others increased such tweeting “dramatically following Musk’s takeover” and do not show signs of slowing down. They found that Twitter hasn’t made progress on bots, which have remained as prevalent and active on Twitter as they were prior to Musk’s tenure.
Musk previously indicated that Twitter’s recommendation algorithms surface less offensive content to people who don’t want to see it.
Keith Burghardt, one of the authors of the paper and a computer scientist at the University of Southern California’s Information Sciences Institute, told CNBC that the deluge of hate speech and other explicit content correlates to the reduction of people working on trust and safety issues and the relaxed content moderation policies.
Musk also said at the WSJ event that “most advertisers” had come back to Twitter.
Louis Jones, a long-time media and advertising executive who now works at the Brand Safety Institute, said it’s not clear how many advertisers have resumed spending but that “many advertisers remain on pause, as Twitter has limited reach compared to some other platforms.”
Jones said many advertisers are waiting to see how levels of “toxicity” and hate speech on Twitter change as the site appears to slant towards more right-wing users and as U.S. elections approach. He said one big challenge for brands is that Musk and Twitter haven’t made clear what they count in their measurements assessing hate speech, spam, scams and bots.
Researchers are calling on Musk to provide data to back up his recent claims.
“More data is critical to really understand whether there is a continuous decrease in either hate speech or bots,” Burghardt said. “That again emphasizes the need for greater transparency and for academics to have a freely available data.”
Show us the data
Getting that data is becoming harder.
Twitter recently started charging companies for access to its application programing interface (API), which allows them to incorporate and analyze Twitter data. The lowest-paid tier costs $42,000 for 50 million tweets.
Imran Ahmed, CEO of the Center for Countering Digital Hate nonprofit, said that because researchers now have “to pay a fortune” to access the API, they’re having to rely on other potential routes to the data.
“Twitter under Elon Musk has been more opaque,” Ahmed said.
He added that Twitter’s search function is less effective than in the past and that view counts, as seen on certain tweets, can suddenly change, making them unstable to use.
“We no longer have any confidence in the accuracy of the data,” Ahmed said.
The CCDH analyzed a series of tweets from the beginning of 2022 through February 28, 2023. It released a report in March analyzing over 1.7 million tweets collected using a data-scraping tool and Twitter’s search function and discovered that tweets mentioning the hateful “grooming” narrative have risen 119% since Musk took over.
That refers to “the false and hateful lie” that the LGBTQ+ community grooms children. The CCDH found that a small number of popular Twitter accounts like Libs of TikTok and Gays Against Groomers have been driving the “hateful ‘grooming’ narrative online.”
The Simon Wiesenthal Center, a Jewish human rights group, continues to find anti-Semitic posts on Twitter. The group recently conducted its 2023 study of digital terrorism and hate on social platforms and graded Twitter a D-, putting it on par with Russia’s VK as the worst in the world for large social networks.
Rabbi Abraham Cooper, associate dean and director of global social action agenda at the center, called on Musk to meet with him to discuss the rise of hate speech on Twitter. He said he has yet to receive a response.
“They need to look at it seriously,” Cooper said. If they don’t, he said, lawmakers are going to be called upon to “do something about it.”
A Tesla Inc. robotaxi on Oltorf Street in Austin, Texas, on June 22, 2025.
Tim Goessman | Bloomberg | Getty Images
Tesla has obtained a permit to operate a ride-hailing service in Arizona, the state’s department of transportation said.
The electric vehicle company applied for a “transportation network company” permit on Nov. 13, and was approved on Monday, ADOT said in an emailed statement. Additional permits will be required before Tesla can operate a robotaxi service in Arizona.
In July, Tesla applied to conduct autonomous vehicle testing and operations in Phoenix, with and without human safety drivers on board. A month earlier, Tesla started a robotaxi pilot in Austin, Texas, with safety valets and remote operators. Tesla also operates a more traditional car service in the San Francisco Bay Area.
Tesla didn’t immediately respond to a request for comment.
Tesla plans to take human safety drivers out of its cars in Austin before the end of this year. The company is aiming to operate a commercial robotaxi service in Phoenix and several other U.S. cities before the end of 2026.
According to the National Highway Traffic Safety Administration’s website, Tesla cars equipped with automated driving systems were involved in seven reported collisions following the launch of the company’s pilot in Texas.
Competitors including Alphabet’s Waymo in the U.S. and Baidu’s Apollo Go in China are way ahead in the nascent robotaxi ride-hailing market. In the Phoenix area, Waymo operates a sizable commercial business, with at least 400 autonomous vehicles, the company previously told CNBC. In May, Waymo said it had surpassed 10 million driverless trips served to riders across the U.S.
Baidu said in an earnings update on Tuesday that its Apollo Go service “provided 3.1 million fully driverless operational rides in the third quarter of 2025,” representing year-over-year growth of 212%.
Musk has been promising that Tesla will “solve” autonomy for years without reaching its goals. The world’s richest person has continued with the lofty pronouncements.
At the company’s 2025 shareholder meeting earlier this month, Musk said the “killer app” for self-driving technology is when people can “text and drive,” or “sleep and drive.”
“Before we allow the car to be driven without paying attention, we need to make sure it’s very safe,” Musk said. “We’re on the cusp of that. I know I’ve said that a few times. We really are at this point.”
Money keeps flowing into artificial intelligence companies but out of AI stocks.
In what looks like — once again — a scenario of the left hand scratching the right, Microsoft and Nvidia will be investing a combined $15 billion into Anthropic, while the OpenAI competitor has committed to buying compute power from its two newest stakeholders. At this point, it seems as if a big proportion of AI news can be summarized as: “Company X invests in Company Y, and Company Y will buy things from Company X.”
Okay, that’s unfair. There are a lot of developments in the AI world that are not about investments but, well, development. Google unveiled the third version of Gemini, its AI model, which Demis Hassabis, CEO of Google’s AI unit DeepMind, said “will be “trading cliché and flattery for genuine insight.” (But I still want an AI chatbot to compliment me on my curiosity when I ask how to cut a pear, so I’m not sure if that’s a pro for me.)
Investors, however, still appear skeptical about AI. Major names such as Nvidia, Amazon and Microsoft tumbled Tuesday stateside, giving the S&P 500 its fourth straight session in the red — the longest decline since August.
And if Nvidia — “the top company within the top industry within the top sector,” as CFRA’s chief investment strategist Sam Stovall puts it — fails to satisfy investors’ expectations when it reports earnings Wednesday, we might be seeing the S&P 500’s slide extend.
Anthropic signs deal with Microsoft and Nvidia. Microsoft announced Tuesday it will invest up to $5 billion in the startup, while Nvidia will put in up to $10 billion. That puts Anthropic’s valuation around $350 billion, according to a source.
Google announces its latest AI model Gemini 3. Alphabet CEO Sundar Pichai said Tuesday it will require “less prompting” for desired answers. The update comes eight months after Google introduced Gemini 2.5, and will be rolled out in the coming weeks.
[PRO] Potentially resilient stocks amid AI slump. There are some global stocks and non-equity assets that could weather the turbulence in U.S. tech names happening recently, strategists told CNBC.
Miffed over Japanese Prime Minister Sanae Takaichi’s comments related to Taiwan, China on Friday advised its citizens against travelling to the country. Japanese tourism-exposed stocks fell in the aftermath of that warning, while experts caution the impact could be more severe over a longer duration.
Takahide Kiuchi, executive economist at Nomura Research Institute, said tensions between the two Asian powers could result in a 1.79 trillion yen drop in Japan’s GDP over the course of one year — a 0.29% decline in the country’s GDP.
Every weekday, the CNBC Investing Club with Jim Cramer releases the Homestretch — an actionable afternoon update, just in time for the last hour of trading on Wall Street. Markets: Stocks continued their recent declines Tuesday as megacap tech lagged on worries about valuations within the artificial intelligence trade. The S & P 500 was on track for its worst losing streak since August as it closed in on its fourth consecutive session of losses. Club stocks Amazon and Microsoft weighed on the market, shedding 4% and 2.7%, respectively, in the afternoon. Club holding Nvidia ‘s 1.5% drop didn’t help sentiment either, going into its highly anticipated earnings report Wednesday evening. The Club also had a busy day of trades. We bought more Home Depot on its post-earnings decline , and sold half of our Disney stake following a disappointing quarter last week. Later in the session, we booked some big profits in Eli Lilly , while adding to our Nike position. The Club also initiated a position in Procter & Gamble , a consumer powerhouse behind household brands like Tide, Crest, and Gillette. Done deal : Salesforce closed its $8.3 billion acquisition of AI-powered data management company Informatica ahead of schedule. The companies had been targeting early next year for completion. “The market didn’t really care for this deal when it was announced in May,” Jeff Marks, director of portfolio analysis for the Club, said Tuesday afternoon, recalling Salesforce shares sinking on reports of the deal and the subsequent announcement a few days later. Marks added that the early completion of the purchase is a “good sign of confidence in the integration that Salesforce expects the deal to be accretive to non-GAAP operating margin and non-GAAP earnings per share one year faster than originally believed.” Despite these positive developments, Marks said Salesforce is still a “show me” story. Salesforce has yet to convince investors that AI doesn’t threaten the software giant’s core business, which operates using a seat-based model. The stock lost more than 1.5% in Tuesday’s trading. Big win: Meta Platforms got a big win Tuesday afternoon in an important antitrust case against the Federal Trade Commission. A federal judge ruled that the FTC did not prove its claims that Meta holds a monopoly in social networking or that the company should not have been allowed to acquire Instagram and WhatsApp back in 2012 and 2014, respectively. The agency, which wanted those two units to be divested, argued that there are no major apps like Facebook and Instagram. The judge, however, said that there are plenty of competitors, citing TikTok and YouTube, and contended that the social media landscape has changed radically since those Meta acquisitions were made over a decade ago. Shares of Club name Meta turned positive late Tuesday. The favorable Meta ruling came 10 weeks after Alphabet’s Google avoided the harshest penalties in the antitrust case it lost last year. Good news: iPhone sales in China surged in October, taking Apple’s dominance in the country’s smartphone market to one in every four phones sold, according to the latest data from Counterpoint Research . Apple last achieved this milestone in 2022. Overall, sales for Apple’s flagship device in China jumped 37% last month from the year prior. Analysts at Counterpoint pointed to solid demand for the iPhone 17, in particular, for the market share gains. All three iPhone 17 variants have outperformed iPhone 16 models in sales, according to Counterpoint, posting mid-to-high double-digit percentage growth from year-earlier levels. The base model of the iPhone 17 continued to grow at the fastest rate. Apple shares were up slightly on Tuesday. Jim Cramer has pounded the table on the new iPhones since the September launch. He previously described its debut as “gigantic” and argued that Apple’s newest devices are “more of a bargain” than past versions. The Club maintains its long-held “own it, don’t trade it” thesis on Apple stock. Up next: Club holding TJX will report quarterly earnings Wednesday morning, along with other retailers like Target and Lowe’s . Then, Nvidia and Palo Alto Networks , both Club names, will release their results after Wednesday’s market close. Investors will also get the minutes from the October Fed meeting at 2 p.m. ET on Wednesday. (See here for a full list of the stocks in Jim Cramer’s Charitable Trust.) As a subscriber to the CNBC Investing Club with Jim Cramer, you will receive a trade alert before Jim makes a trade. Jim waits 45 minutes after sending a trade alert before buying or selling a stock in his charitable trust’s portfolio. If Jim has talked about a stock on CNBC TV, he waits 72 hours after issuing the trade alert before executing the trade. THE ABOVE INVESTING CLUB INFORMATION IS SUBJECT TO OUR TERMS AND CONDITIONS AND PRIVACY POLICY , TOGETHER WITH OUR DISCLAIMER . NO FIDUCIARY OBLIGATION OR DUTY EXISTS, OR IS CREATED, BY VIRTUE OF YOUR RECEIPT OF ANY INFORMATION PROVIDED IN CONNECTION WITH THE INVESTING CLUB. NO SPECIFIC OUTCOME OR PROFIT IS GUARANTEED.