TheChek News Flow

The artificial-intelligence giant is working with bankers at Goldman Sachs and Morgan Stanley

ChatGPT-maker OpenAI has been working with bankers to prepare to file for an initial public offering in the coming days or weeks, according to people familiar with the matter.

Bankers at firms including Goldman Sachs GS 4.39%increase; green up pointing triangle and Morgan Stanley MS 3.87%increase; green up pointing triangle have been helping the artificial-intelligence giant on a draft IPO prospectus it plans to file confidentially with regulators soon, possibly as early as Friday, some of the people said.

The goal is for the artificial intelligence giant led by Sam Altman to be ready to go public as early as September, some of the people said. The plans remain fluid and could still change, they cautioned.

The company cleared a major roadblock to its IPO earlier this week, when it won a legal fight with its co-founder turned Altman foe Elon Musk. Musk, whose rocket-maker SpaceX is expected to unveil IPO paperwork Wednesday ahead of a potential June offering, has said he plans to appeal the verdict.

OpenAI will have to overcome a host of other challenges, namely concerns about whether the company will be able to generate enough revenue to support its giant data-center spending commitments.

OpenAI’s lead in the AI race has recently come under threat from Anthropic, which has grown faster than its rival in recent months thanks to the fast adoption of its software tools across the workforce. OpenAI is in the middle of a major strategy pivot to try and catch up.

Lorna Hajdini countersued on Tuesday after being named in harassment lawsuit that went viral on social media

Lorna Hajdini, the JPMorgan investment banker who was accused of sexual assault, countersued on Tuesday, saying the AI-generated memes and coverage of the lurid allegations have made her life “a daily, living nightmare.”

In a suit against her former colleague, Chirayu Rana, Hajdini said his allegations were entirely fabricated, have caused her mental harm and resulted in her losing a volunteering job. She is suing him for defamation and “malicious prosecution.”

“Plaintiff deliberately crafted a salacious narrative designed to achieve maximum press coverage and inflict maximum pain; unfortunately, it has succeeded despite its falsity,” the countersuit said. “Plaintiff’s false, malicious, and bad faith statements have wreaked havoc on Ms. Hajdini’s life.”

She also claimed that Rana’s attacks on her were part of a “common pattern and scheme” he had advanced at other employers, though she didn’t name a specific firm. He has worked at a litany of well-known firms on Wall Street, including Morgan Stanley, Carlyle Group and an affiliate of Apollo Global Management.

Rana had originally filed a lawsuit on April 27 under the pseudonym “John Doe,” alleging that Hajdini had sexually harassed and assaulted him for months. The graphic detail in its allegations and quotations that it claimed Hajdini said went viral on social media, including “If you don’t f—me soon, I’m going to ruin you…Never forget, I f—ing own you,” and “If you don’t f—my brains out tonight, I’m going to sabotage your promotion.”

Daniel Kaiser, a lawyer for Rana, didn’t immediately reply to a request for comment. He has previously said that “we look forward to the truth and facts coming to light through court proceedings.”

JPMorgan said in a statement that “we fully support Lorna and her right to defend herself and protect her reputation. As we’ve said from the outset, we don’t believe the allegations against her or the firm have merit.”

Hajdini’s lawyers had denied the allegations from the start. JPMorgan has said the bank conducted an internal investigation after Rana filed a complaint and found them not credible. The bank still offered him a $1 million payout, but Rana rejected the settlement and went public, The Wall Street Journal reported.

Rana’s allegations were a hot topic across Wall Street and social media for weeks, as many observers doubted they were truthful. In her countersuit, Hajdini says the public reaction took a toll on her life and she has received threats to her personal safety.

She started at JPMorgan in 2011 and had never been the subject of a harassment complaint, the suit said. Rana’s lawsuit has resulted in lasting damage to her career, Hajdini’s suit said.

“Ms. Hajdini and members of her family have been mocked, ridiculed, and harassed around the clock, with Ms. Hajdini serving as the ongoing focal point of countless jokes, memes, and AI-generated images and videos of a persistently vile, degrading, and sexual nature—all a direct consequence of Plaintiff’s lies,” the suit said.

New ad formats will run in AI Mode and standard search but not in Gemini (yet)

Google has begun testing several new ad formats in both standard search results and its AI Mode as the race to turn artificial intelligence into ad revenue picks up.

Researchers predict that Instagram and Facebook parent Meta Platforms may soon surpass Google as the world’s largest seller of digital advertising thanks in large part to the increased automation of campaign creation and distribution. OpenAI meanwhile has been testing ads in ChatGPT since February.

But the companies are trying to be careful not to alienate generative AI users, who may have different expectations for responses to their prompts than they do for simpler search terms. OpenAI rival Anthropic has already attacked the ChatGPT ads test with a Super Bowl commercial.

“In a world where AI provides a direct and conversational response, will users still find value in ads?” said Shashi Thakur, vice president and general manager of Google search ads and ads on Google experiences. “Our answer is, it’s definitely yes, but you have to rethink how to approach ads and what an ad is.”

The U.S. rollout comes one day after Google introduced other AI products such as a personal agent called Gemini Spark and the video-creation tool Gemini Omni.

AI-generated ads for brands and products will now appear below the responses to some users’ prompts in AI Mode, according to Thakur, who oversaw the development of the new products. The company’s so-called “direct offers” ads for discounted products will also expand from AI Mode to standard search results.

For now, the ads will not appear in Gemini, the company’s large language model app. But they may represent the next step in that direction, according to Michael Nathanson, senior research analyst at the research firm MoffettNathanson.

“They just have been deliberately trying to get the right AI products in front of you,” Nathanson said, referring to Google’s expanding suite of search tools. “And now comes the monetization phase of it all.”

Google at the moment is focused on AI Mode, Chief Business Officer Philipp Schindler said on the company’s most recent earnings call when Nathanson asked about bringing ads to Gemini.

“But it’s fair to say that we really believe a format that works well in AI mode would transfer successfully to Gemini app,” Schindler said.

Google has expanded the direct offers format since its introduction earlier this year to include hotels and other travel providers as well as a “bundle” feature combining several deals from the same retailer in a single ad.

For its new, more elaborate “conversational discovery” format, Google uses Gemini to create ads based on users’ prompts by drawing text and imagery from brands’ websites and marketing materials.

More specific “highlighted answer” ads include product explanations also generated by Gemini from a company’s own materials.

Each ad format will be labeled with the word “sponsored,” a Google spokeswoman said.

Google has also begun testing new formats in traditional search results. AI-powered shopping ads again explain products, while a “business agent for leads” tool lets users ask questions about a business or product within an ad itself.

The ads are designed to reach consumers at different points in the progression from their initial searches to ultimately making purchases, Thakur said. Conversational ads introduce them to brands they may not otherwise know, for example, while direct offers may show up when users ask to see products at certain price points, he said.

Gemini’s reliance on a brand’s own content when generating the ads may reinforce marketers’ growing focus on designing that material specifically for LLM searches, sometimes with the help of proliferating “answer engine optimization” startups. Advertisers will not have full control over the final copy or imagery, but they can use a tool called AI Brief to emphasize or remove certain phrases. A brand may, for example, tell Gemini not to describe its products as “low cost,” Google said.

Google will protect the ads against AI hallucinations, or false information, with “very strong guardrails,” said Thakur, who declined to elaborate.

Meta’s sustained growth and OpenAI’s entry into the ad market leave Google little choice but to quickly introduce ads in Gemini, said Paul Armstrong, founder of tech and data consulting firm TBD Group and author of the newsletter “What Did Google Do This Week?”

At the same time, the company must move cautiously to avoid incurring the wrath of ad-averse users who see LLM responses as a final destination for their queries rather than a guide to other online resources, said Armstrong.

“Google spent 20 years training users to tolerate ads because search still felt useful,” he said. “Once ads start blending into [AI] generated responses, users stop feeling like they are searching and start feeling like they are being handled.”