A top Wall Street analyst used ChatGPT and Gemini to assess Google risks. He came away impressed.

Business Insider

2 days ago

In a striking twist to a major legal showdown between the US government and one of the world's most powerful tech companies, Wall Street analysts aren't just turning to lawyers — they're consulting artificial intelligence.
And not just any AI: Evercore ISI's top tech analyst Mark Mahaney tapped Google's own Gemini AI chatbot, along with OpenAI's ChatGPT, to assess what could happen next in the Department of Justice's big case against Google's Search business.
With Judge Amit Mehta expected to rule later this summer, Mahaney published a research note that offers a rare look at how investors use generative AI as a strategic analysis tool for watching one of the most consequential antitrust cases of the 21st century.
AI-powered legal analysis on Wall Street
Mahaney and his team gave OpenAI's o3 and Gemini 2.5 Pro transcripts from the final day of court proceedings — more than eight hours of closing arguments — and asked the chatbots two key questions: What were the main takeaways? And what remedies will Judge Mehta most likely impose?
The result? "Pretty darn impressive," Mahaney wrote.
Within 17 seconds, the models generated concise, reasoned summaries and remedy predictions. The output from both AI tools aligned closely with the conclusions of legal experts consulted by Evercore, suggesting the technology may be ready for prime time as a research aid, even in complex regulatory situations, Mahaney wrote.
"GenAI reasoning may become a table-stakes capability... and certainly can help achieve improved productivity goals for users, enterprises, consumers and Wall Street Internet Analysts!," he wrote.
I emailed to ask Mahaney whether he's trying to put himself out of a job.
Mahaney replied, "😊".
AI's predictions for Google
Both ChatGPT and Gemini forecast that Judge Mehta will likely impose a slate of tough behavioral remedies and stop short of ordering a huge structural breakup of Google's empire. Their projected outcomes included:
A highly probable ban on default Search payments (ChatGPT was more sure of this, while Gemini thought a complete, permanent ban was unlikely. Shocker!)
It's unlikely that Google will be forced to sell its Chrome browser business, though Judge Mehta might keep this in reserve if other remedies don't spark more competition. (Both chatbots aligned on this prediction pretty closely.)
Likely contingent limits on general Search revenue-share payments.
Possible contingent requirements on data sharing.
ChatGPT's detailed reasoning suggested Judge Mehta is interested in avoiding collateral damage to the tech ecosystem while still targeting the "monopoly flywheel" Google uses to maintain dominance.
Gemini's forecast echoed this logic, emphasizing Mehta's focus on balancing enforcement with market disruption risk and the avoidance of a "massive structural breakup."
Why this matters for Google and Wall Street
Mahaney's use of AI wasn't about replacing expert human judgment, but rather augmenting it. He and his team had already reviewed transcripts and interviewed legal experts. However, the speed and coherence with which ChatGPT and Gemini processed complex legal dialogue and arrived at similar conclusions gave Mahaney confidence that GenAI tools can assist in forecasting high-stakes regulatory events.
The case, which stems from Google's multibillion-dollar payments to companies like Apple to maintain default status on browsers and smartphones, could upend how the $2 trillion internet giant competes in search and beyond.
For investors, the insight that AI models can accurately mirror expert expectations in such nuanced legal territory may prove as significant as the case itself.
As Judge Mehta finalizes his ruling, the courtroom might still be human, but Wall Street's analysis of it is increasingly AI-assisted. Welcome to the era of algorithmic legal foresight.