Fintech

NASA and Google Test AI ‘Space Doctor’ | PYMNTS.com

Published

1 week ago

August 23, 2025

Star Trek’s fictional Dr. “Bones” McCoy may have had a tricorder, but NASA astronauts could soon have something almost as futuristic and real: An artificial intelligence (AI)-powered medical assistant that can diagnose and help treat illnesses millions of miles from Earth.

Google and NASA are testing the “Crew Medical Officer Digital Assistant,” an AI-powered system to help astronauts with their healthcare needs during extended space missions to the moon, Mars or even farther out, according to a Aug. 7 Google blog post. Such long-range missions can experience a lag in communications with those on Earth.

“Currently, low Earth orbit missions such as those to the International Space Station have frequent and relatively robust connectivity to allow for close collaboration with ground-based support teams,” David Cruley, customer engineer of Google Cloud Platform, told PYMNTS. “However, as distance from the Earth increases, so will latency and communication gaps.”

The communication lag can be as long as 22 minutes one-way on a mission to Mars, according to an April report from NASA, “Assessment of the State of Communication Delay Research in Preparation for Missions Beyond Low Earth Orbit.” Missions to the moon can experience latencies of three to 14 seconds one-way, but Mars missions can have up to a 22-minute one-way delay or 44 minutes round-trip.

Since every second counts in a medical emergency, the AI doctor steps into the gap while the crew waits for connectivity to catch up.

While there will be a human medical officer on board the spacecraft, Cruley said more information and support might be needed in diagnosing medical issues, especially for longer-range space missions. Moreover, if the medical officer is the one with the health problem, the rest of the crew would need help from the AI-powered doctor.

This development comes at a time when NASA is preparing for a new stage of space missions, starting with the Artemis campaign to set up its first long-term presence on the moon. The space agency said on its website that an exploration of the moon will yield scientific discovery, technology advancement and lessons on living and working “on another world as we prepare for human missions to Mars.”

Google and NASA said on the Google post they are currently in the initial phases of testing the AI space doctor. It has undergone preliminary trials in which a panel consisting of doctors and an astronaut evaluated its performance in simulated medical scenarios, Cruley told PYMNTS.

“The next step is to continue testing the model with medical doctors to enhance its accuracy and reliability,” he added.

The AI doctor is trained on spaceflight data and uses natural language processing and machine learning to give astronauts timely, data-driven guidance for managing medical issues during missions.

The model’s performance is measured using what’s called the Objective Structured Clinical Examination, according to Google. This exam is used to assess the skills of medical students and professionals, according to the National Institutes of Health.

Initial results show the AI could deliver “reliable” diagnoses based on reported symptoms, according to Google. The company and NASA are now working with medical experts to refine the system’s accuracy and decision-making.

Cruley says there’s no launch date yet for the AI space doctor. But while the project’s primary mission is to support astronauts, its potential reach is much broader.

“The idea of an AI Digital Health Assistant is portable to Earth-based applications,” Cruley said. “Lessons learned could be applied to providing quality medical care in remote or underserved areas with limited access to healthcare professionals.”

Fintech

European Commission to Impose ‘Modest’ Penalties in Google AdTech Case | PYMNTS.com

Published

19 hours ago

August 31, 2025

Hrishi

When the European Commission announces the penalties in an antitrust case involving Google’s AdTech business in the coming weeks, it will reportedly order the company to pay a “modest” fine and will not require it to sell part of its AdTech business.

The fine is likely to be less than the 4.3 billion euros the Commission ordered Google to pay in another case in 2018, Reuters reported Friday (Aug. 29), citing unnamed sources.

Asked about the report by Reuters, the Commission decline to comment, while Google pointed to a 2023 blog post in which it criticized the regulator’s interpretation of the AdTech sector, according to the report.

The report attributed the imposition of what it described as modest penalties to the approach of the new EU antitrust chief, Teresa Ribera, who it said focuses on getting companies to end anti-competitive practices rather than punishing them with big fines.

In the AdTech case, the European Commission accused Google of abusing its dominance in the online AdTech industry since 2014. The regulator alleged that the company had wielded its market power on both sides of the supply chain by showing favoritism toward its own ad exchange, AdX, in matching auctions.

Google argued that serving both advertisers and publishers is a common industry practice, competitors operate similar AdTech businesses catering to both sides of the market, and integrated technology stacks facilitate high-quality connections between advertisers and publishers.

According to Friday’s Reuters report, advertising revenue accounted for 75.6% of Google’s total revenue in 2024.

Google also faces pressure in the United States, where it is in a legal battle with the Justice Department, and a federal judge determined that the company unlawfully maintained monopolies in critical areas of the online ad industry.

It was reported in May that the Justice Department called for Google to divest some key components of its digital advertising business, including its AdX marketplace and its DFP ad-serving platform.

The Justice Department argued at the time in a court filing that such divestitures were essential to dismantle Google’s dominance and restore competition in the markets for ad exchanges and publisher ad servers.

Fintech

AI Model Betting Is the New Fantasy Football | PYMNTS.com

Published

2 days ago

August 30, 2025

Hrishi

Sports gambling has DraftKings. Political junkies have PredictIt. And now the world’s nerdiest corner — the artificial intelligence (AI) scene — has its own set of bettors, where people wager actual money on whether Google’s Gemini will dunk on OpenAI’s GPT-5 this month.

Forget fantasy football, this is fantasy machine learning.

People are placing their bets on markets like Kalshi, where they can trade on the outcome of real-world events, everything from when Taylor Swift and Travis Kelce will wed to whether Google will break up and, of course, the AI model race. Kalshi saw 10 times the volume on AI trades compared with the start of the year, according to The Wall Street Journal.

“Kalshi’s markets are extremely efficient and serve as a source of truth on the likelihood of all events, including AI model progress,” Jack Such, a spokesperson for the company, told PYMNTS.

So who’s winning based on the bets?

“Gemini is the current market leader for ‘Best AI’ by the end of 2025,” Such said.

On Kalshi, Gemini shows a 58% probability of winning, compared with 19% for ChatGPT, as of midmorning on Thursday (Aug. 28). The third spot goes to Grok, at 17%. Claude, which Menlo Ventures said was the enterprise favorite, clocks in at 2% and tied with Meta’s Llama.

The total trading volume reached $8.1 million.

Kalshi is just one of several platforms that are pivoting to AI trading.

Other players include Polymarket, an offshore crypto-based prediction market, Manifold Markets, Metaculus and other sites.

Which AI model is winning globally? The AI community abroad agrees with the Americans. Polymarket, which is only available to non-Americans, predicts that Google will win by year’s end, at 66% odds. OpenAI comes in at 16% probability while xAI is third with 14%.

See also: Crypto Firms Grapple with Bank-Like Risks, Without the Regulation

Betting on Anything in the World

How Kalshi works: People bet yes or no on outcomes in the real world, such as whether the Fed will cut short-term interest rates by 25 basis points in September (76% probability it will) or who will win the Nobel Peace Prize (Yulia Navalnaya, widow of Russian opposition leader Alexei, is at 20% probability and Donald Trump at 10%).

Bettors buy contracts tied to these outcomes at prices between a penny and 99 cents, settling at $1. For example, if a bettor buys a contract for 40 cents and they guess the outcome correctly, they are paid $1. If they guessed wrong, they lose their 40 cents. Bettors can also sell the contract before there’s an outcome if they see the contract price go up or down.

In a tie, the default would be a negative outcome.

For AI models, results will be determined by the rankings on the LMSYS Chatbot Arena Leaderboard at the end of the year, according to Kalshi.

Currently, the leaderboard shows Gemini 2.5 pro slightly ahead of GPT-5 for first place, followed closely by Claude Opus.

One Kalshi user noticed a tie earlier this month and wrote, “it’s a tight race on LMArena. I don’t understand why the spread is so drastic. Huge earnings for people who bet on GPT with decent odds that it will flip.”

Kalshi began allowing bettors to trade on AI models in 2023. “We were confident in consumer demand for these products because of the rapid growth of AI and the economic and political consequences of model progress,” Such said.

Kalshi said the U.S. Commodity Futures Trading Commission regulates it as a financial exchange for trading futures, swaps and options on commodities.

“Understanding that Kalshi is regulated … helps reassure users that they are engaging with a platform that adheres to the highest standards of operation and accountability,” the company says on its site.

Here’s to hoping the bet will be good.

Read more:

Cross-Border Payments Platform Wise Proposes Moving Stock Market Listing to US

Ripple Prepares to Bring Stablecoin to Global Exchanges

FTX Drops Plans to Resurrect Fallen Crypto Exchange

Fintech

Chinese Money Laundering Networks Flagged as ‘Severe Threat’ by FinCEN | PYMNTS.com

Published

3 days ago

August 29, 2025

Hrishi

The Treasury Department’s Financial Crimes Enforcement Network (FinCEN) has issued a warning that Chinese money laundering networks are now among the most significant threats to the U.S. financial system, fueling operations of Mexico’s most powerful drug cartels.

In a new advisory released Thursday (August 28), FinCEN said groups including the Sinaloa Cartel and the Jalisco New Generation Cartel are increasingly using Chinese money laundering networks to clean billions in drug proceeds. The networks have become effective partners because they can move cash quickly, absorb losses and leverage demand from Chinese nationals seeking to bypass Beijing’s strict currency controls. By pairing cartel dollars with Chinese demand for U.S. currency, these networks have created what FinCEN called a “mutualistic relationship” that strengthens both sides.

The 15-page report outlines how these networks operate. They rely on informal value transfer systems, “mirror” transactions and trade-based money laundering schemes to shift funds across borders without physically moving cash. Some recruit Chinese students in the U.S. to act as money mules, depositing cartel cash into U.S. bank accounts and disguising the flows as tuition or living expenses. Others purchase electronics, luxury goods, and real estate with illicit funds, later reselling or exporting those assets to create the appearance of legitimate commerce.

“Chinese money laundering networks are professional money launderers that play a vital role in laundering the cartels’ drug proceeds in the United States,” the advisory said. “This is due to the speed and effectiveness of their operations, as well as their willingness to absorb financial losses and assume risks for the cartels and other clients.”

The advisory also lists “red flag” indicators for financial institutions, ranging from students depositing unusually large sums of cash to small businesses with outsized luxury goods transactions. FinCEN urged banks to reference the identifier “CMLN-2025-A003” when filing suspicious activity reports, and to pay special attention to transactions tied to unlicensed money service businesses, informal transfer systems and trade-based schemes.

FinCEN’s move follows a January executive order that designated several cartels as foreign terrorist organizations, underscoring the administration’s concern that cartel financing now poses not only a criminal but also a national security threat. The advisory ends with a reminder that financial institutions are on the front line of detecting suspicious flows and that voluntary information sharing under existing safe harbors can help expose networks that otherwise thrive in the shadows.