AI Shows Promise in Bridging Business Divides, Experts Say

Google Deepmind, AI, Math Olympiad

AI software that crafts consensus statements from opposing viewpoints could one day be a tool for smoothing corporate negotiations and stakeholder disputes.

A new breed of artificial intelligence system can analyze conflicting positions and generate balanced group statements that capture majority and minority perspectives. This could transform how businesses handle everything from labor talks to merger discussions. Researchers and business consultants say AI could help parties find common ground faster than traditional mediation.

“We have to work together, we have to collaborate,” Hanne Wulp, executive consultant and founder of Communication Wise told PYMNTS. “These hardened, far out-of-the-middleground perceptions don’t come in handy. There won’t be many others to collaborate with. When AI-driven mediation can soften, or tweak, that lens slightly, just by gathering and framing perceptions in a neutral, non-confrontational/collaborative way, it can enhance collaboration.”

Study Shows Promise in Consensus Building Tool

The new AI tool developed by Google DeepMind shows promise in bridging ideological divides through group discussions. In a study published in Science, researchers found that their “Habermas Machine” — based on the Chinchilla language model — effectively synthesized opposing viewpoints into consensus statements. Testing with 439 UK residents revealed that 56% preferred AI-generated summaries over human mediators. The system could improve citizens’ assemblies and public policy discussions by creating more balanced, representative statements — or even have commercial implications. 

“Group statements generated by AI can integrate the needs, opinions, and cultural backgrounds of different consumers,” Alex Li, Founder of AI company StudyX, told PYMNTS. “This inclusive marketing strategy can resonate with more consumers, enhance brand appeal, and finally influence consumer behavior, making them inclined to choose products that align with their values.”

AI systems are increasingly helping businesses reach consensus in complex commercial decisions. OpenAI’s Swarm Framework allows multiple AI agents to work together, streamlining decision-making. Google’s Gemini models enhance negotiation capabilities, helping companies align transaction interests. IBM’s Watson assists supply chain management by analyzing data from different stakeholders, leading to agreed-upon solutions for sourcing and logistics. Additionally, platforms like Pactum automate contract negotiations, ensuring fair terms for all parties. 

Skeptics Abound

Not everyone’s a fan of AI taking charge of sending out group statements. Michael Taylor, CEO of SchellingPoint, which manages what he describes as “the world’s largest database of real-time group decisions” with over 9 million data points, told PYMNTS he’s skeptical about AI-generated group consensus. 

Taylor explains that when groups first discuss a shared topic, “17% of their opinions are like-minded, and 83% are non-likeminded.” Using a framework based on Harvard Professor Chris Argyris’s work, his organization analyzes why people agree or disagree.

He identifies key patterns in group disagreement: “30% of the time” differences stem from varying access to information, while “65% of the time” disputes arise from different interpretations of terminology. He argues that both cases can be resolved through understanding rather than compromise.

“Replacing the reconciliation of non-aligned opinions with suggested group statements to gain consensus would significantly compromise the accuracy and integrity of the strategies, policies, decisions and changes these groups are forming,” Taylor warns.

Instead, SchellingPoint has developed an AI system that analyzes group thinking patterns and helps determine accurate conclusions rather than seeking consensus for consensus’s sake.

“AI should be used iteratively at the individual level not for consensus, but rather to counter internal bias, poorly actuated mental heuristics, and undiscovered cognitive dissonance,” Christopher Kaufman, a professor of Business and Leadership Studies at Westcliff University, told PYMNTS. “Then each person, after their own iteration of bias and dissonance discovery, can then use AI to present their own new formatted concepts. And then see if their new ideas are adopted by human consensus.”


Mastercard and Feedzai Team to Fight AI-Powered Scams

Feedzai, Trust Payments, risk management

Mastercard has teamed with financial crime prevention company Feedzai to help financial institutions prevent scams.

“As payments continue to evolve, fraudsters are increasingly using AI to scam consumers,” the companies said in a news release Tuesday (Feb. 18). “This cost more than $1 trillion last year, with more than 50% of consumers saying they had encountered a scam at least once a week.”

To that end, Mastercard will leverage Feedzai’s fraud platform, available in more than 90 countries, to deploy its Consumer Fraud Risk (CFR) solution to customers across many key markets around the world.

According to the release, CFR provides sending and receiving financial institutions in account-to-account payment transactions with intelligence to spot and prevent scams in real time. Since going live in the U.K. in 2023, the value of authorized push payment (APP) scams has fallen by more than 12% in that country.

Feedzai’s AI-native Financial Crime Prevention Platform, the release adds, is used by leading financial institutions to protect more than a billion consumers worldwide, and upwards of $8 trillion in transactions each year.

“With more than half the world’s population affected, the scale of scam fraud is not only having a devastating impact on consumers, but also surpassing the GDP of many individual economies,” said Johan Gerber, executive vice president, head of Security Solutions at Mastercard. “Together with Feedzai’s global platform we will scale our first-of-its-kind scams solution to more markets, helping more financial institutions combat financial crime faster than before.”

PYMNTS Intelligence research has found that scams became the leading form of fraud last year, ahead of digital payment fraud. The share of scam-related fraud rose by 56%, and financial losses from scams increased 121%. Scams now make up 23% of all fraudulent transactions, with relationship/trust and product/service scams generating the most losses. 

“These scams manipulate individuals into authorizing fraudulent transactions, often using deceptive tactics,” PYMNTS wrote. “Additionally, fraud involving compromised credentials, where individuals are tricked into revealing account details, is also on the rise.”

Additional research by PYMNTS Intelligence has found that many financial institutions (FIs) are turning to artificial intelligence (AI) and machine learning (ML) technologies to detect fraud.

“These technologies analyze large amounts of transaction data in real time, spotting suspicious activity and preventing fraudulent payments before they are completed,” PYMNTS wrote late last year.

The research found that 71% of FIs are using AI and ML for fraud detection, up from 66% in 2023. These technologies can identify anomalies and patterns that might not be detected by human analysts, allowing banks to perform quicker decision-making.