Connect with us

Business

There’s a Looming AI Data Shortage. Google Researchers Have a New Fix.

Published

on


Google DeepMind researchers have an idea for how to solve the AI data drought, and it might involve your Social Security number.

The large language models powering AI require vast amounts of training data pulled from webpages, books, and other sources. When it comes to text specifically, the amount of data on the web considered fair game for training AI models is being scraped faster than new data is being created.

However, a large portion of the data isn’t used because it’s deemed toxic, inaccurate, or it contains personally identifiable information.

In a newly published paper, a group of Google DeepMind researchers claim to have found a way to clean up this data and make it usable for training, which they claim could be a “powerful tool” for scaling up frontier models.

They refer to the idea as Generative Data Refinement, or GDR. The method uses pretrained generative models to rewrite the unusable data, effectively purifying it so it can be safely trained on. It’s not clear if this is a technique Google is using for its Gemini models.

Minqi Jiang, one of the paper’s researchers who has since left the company to Meta, told Business Insider that a lot of AI labs are leaving usable training data on the table because it’s intermingled with bad data. For example, if there’s a document on the web that contains something considered unusable, such as someone’s phone number or an incorrect fact, labs will often discard the entire thing.

“So you essentially lose all those tokens inside of that document, even if it was a small single line that contained some personally identifying information,” said Jiang. Tokens are the units of data, processed by AI, which make up words within text.

The authors give an example of raw data that included someone’s Social Security number or information that may soon be out of date (“the incoming CEO is…”). In these instances, the GDR would swap or remove the numbers, ignore the information that risks becoming obsolete, and retain the remainder of usable data.

The paper was written more than a year ago and was only published this month. A Google DeepMind spokesperson did not respond to a request for comment about whether the researcher’s work was being applied to the company’s AI models.

The authors’ findings could prove helpful for labs as the usable well of data runs dry. They cite a research paper from 2022 that predicted AI models could soak up all the human-generated text between 2026 and 2032. This prediction was based upon the amount of indexed web data, using statistics from Common Crawl, a project that continuously scrapes web pages and makes them openly available for AI labs to use.

For the GDR paper, the researchers performed a proof of concept by taking over one million lines of code and having human expert labelers annotate the data line by line. They then compared the results with the GDR method.

“It completely crushes the existing industry solutions being used for this kind of stuff,” said Jiang.

The authors also said their method is better than the use of synthetic data (data generated by AI models for the purpose of training themselves or other models), which has been a topic of exploration among AI labs. However, using synthetic data can degrade the quality of model output and, in some cases, lead to “model collapse.”

The authors compared the GDR data against synthetic data created by an LLM and discovered that their approach created a better dataset for training AI models.

They also said further testing could be conducted on other complicated types of data considered a no-go, such as copyrighted materials and personal data that is inferred across multiple documents rather than explicitly spelled out.

The paper has not been peer reviewed, said Jiang, adding that this is common in the tech industry and that all papers are reviewed internally.

The researchers only tested GDR on text and coding. Jiang said that it could also be tested on other modalities, such as video and audio. However, given the rate at which new videos are generated each day, they’re still providing a firehose of data for AI to train on.

“With video, you’re just going to have a lot more of it, just because there’s a constant stream of millions of hours of video generated each day,” said Jiang. “So I do think, going across new modalities beyond text, video, and images, we’re going to unlock a lot more data.”

Have something to share? Contact this reporter via email at hlangley@businessinsider.com or Signal at 628-228-1836. Use a personal email address and a non-work device; here’s our guide to sharing information securely.





Source link

Continue Reading
Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *

Business

Former xAI CFO Named OpenAI’s New Business Finance Officer

Published

on


OpenAI has hired Mike Liberatore, the former chief financial officer at Elon Musk’s AI company xAI, CNBC reported on September 16. 

Liberatore’s LinkedIn profile lists his current role as the business finance officer at OpenAI. His tenure at xAI lasted merely four months, and previously, he worked as the vice president of finance and corporate development at Airbnb. 

The report added that Liberatore will report to OpenAI’s current CFO, Sarah Friar, and will work with co-founder Greg Brockman’s team, which manages the contracts and capital behind the company’s compute strategy. 

According to The Wall Street Journal’s report, Liberatore was involved with xAI’s funding efforts, including a $5 billion debt sale in June. He also oversaw xAI’s data centre expansion in Memphis, Tennessee, in the United States. The reasons for his departure remain unknown. 

Liberatore is an addition to the list of recent high-profile departures from xAI. Last month, Robert Keele, who was the general counsel at the company, announced his departure, stating that there were differences between his worldviews and Musk’s. 

The WSJ report also added that Raghu Rao, a senior lawyer overseeing the commercial and legal affairs for the company, left around the same time. 

Furthermore, Igor Babuschkin, the co-founder of the company, also announced last month that he was leaving xAI to start his own venture capital firm. 

That being said, Liberatore’s appointment at OpenAI comes at a time when the company has announced significant structural changes. 

OpenAI recently announced that its nonprofit division will now be ‘paired’ with a stake in its public benefit corporation (PBC), valued at over $100 billion. The company also announced it has signed a memorandum of understanding with Microsoft to transform its for-profit arm into a PBC. This structural change was initially announced by OpenAI in May. 



Source link

Continue Reading

Business

Google Advisor Explains Why ESG-Led AI Is Essential For Business Resilience In The Future Of Work

Published

on




This article is based on the
Future of Work Podcast episode “Why AI and ESG Must Evolve Together to Protect the Future of Work” with Kate O’Neill. Click here to listen to the entire episode.

In the rush to innovate, are today’s leaders forgetting why they started? 

Advertisements

Businesses chasing AI without aligning to human-centered metrics risk building beautiful systems that fail spectacularly.

In a recent episode of The Future of Work® Podcast, Kate O’Neill, CEO of KO Insights and a seasoned digital transformation strategist, delivered a critical message to today’s business leaders: you must stop chasing metrics in isolation and start thinking in terms of ecosystems. 

Advertisements

Nexudus - Tech Stack Lovers

As AI becomes an increasingly central part of how organizations operate, leaders face a choice: retrofit outdated success models to new technologies, or reimagine the system altogether through the lens of purpose, resilience, and human flourishing.

With a career advising clients as varied as Google, McDonald’s, and the United Nations, O’Neill isn’t a futurist just making vague predictions. She’s a strategist with a clear framework and a call to action to solve AI integration problems: align artificial intelligence initiatives with Environmental, Social, and Governance (ESG) principles — not in name only, but in measurable, mission-driven ways that track real-world outcomes.

“I think ESG as a concept is valid. It’s not the principles that are wrong. It’s that we’ve been measuring the wrong things,” she said during the podcast conversation. 

This insight forms the cornerstone of O’Neill’s approach. In a world captivated by AI’s predictive capabilities and automation potential, organizations often overlook the encompassing impact of their decisions. 

Advertisements

Maximize Flexible Space Revenue

Are these technologies improving lives? Are they regenerating ecosystems — social or environmental — rather than extracting from them? Too often, she explains, companies confuse compliance with progress, chasing ESG as a branding exercise instead of a structural transformation.

This critique is not about abandoning ESG or digital transformation. Quite the opposite. It’s about evolving both.

From Checklists to Systems Thinking

The past decade has seen ESG reporting become a staple of corporate responsibility efforts. But O’Neill points out a flaw: ESG frameworks often push businesses to focus on standardized inputs and outputs rather than actual impact

These rubrics, while helpful for consistency, can fail to reflect the lived experience of people and communities affected by a company’s operations.

Instead, she argues for aligning with the United Nations Sustainable Development Goals (SDGs), a framework of 17 interrelated goals with actionable metrics designed to improve life for all — not just shareholders.

To her, that’s a better approach, as most businesses are doing something that could be furthering the SDGs, but they just don’t necessarily realize it.

From water access and infrastructure to gender equality and education, the SDGs provide a nuanced, flexible way for companies to identify where their operations already intersect with meaningful societal progress. 

More importantly, they allow companies to evolve those operations in a direction that’s measurable, values-aligned, and resilient.

Advertisements

2025 GWA Conference

Making ESG Real in the Age of AI

AI technologies are tools that mirror the systems they’re built within. When integrated blindly, AI can amplify inequities and environmental damage. But when aligned with well-defined social goals, it can act as a force multiplier for good.

Consider how companies often rush to replace human labor with AI in the name of efficiency. O’Neill challenges this logic, not just from a social justice perspective but from a business strategy standpoint. In many cases, this kind of substitution overlooks deeper ESG implications — regional job displacement, lost organizational knowledge, reduced resilience in the face of uncertainty.

“Additive” use of AI, she argues, is far more effective than “replacing” strategies. Enhancing human capability, rather than removing it, yields more sustainable organizations.

This philosophy stems from a fundamental distinction O’Neill highlights: the difference between sense-making and prediction.

Advertisements

Yardi Kube automates flex and coworking operations

Humans interpret, synthesize, and apply judgment. Machines, even the most advanced AI, rely on data and probability. One of her favorite analogies comes from healthcare: a doctor can hear the emotional nuance in a teenager’s “I’m fine” — something no large language model can reliably decode today. 

In complex systems — like health, education, or public infrastructure — nuance matters.

A Fast-Changing Landscape Needs Slow, Strategic Thinking

Much of the anxiety among today’s executives comes from the pace of change. Technology is moving faster than ever, and leaders are under pressure to act quickly or risk irrelevance. But as O’Neill notes, movement alone isn’t enough. Strategic motion — guided by values and grounded in measurable, ecosystem-wide outcomes — is what will separate resilient organizations from fragile ones.

The goal is progress, not perfection, and that progress requires recognizing the trade-offs embedded in every transformation decision.

Advertisements

2025 GWA Conference

We are already seeing early-stage consequences: water-intensive AI data centers straining local ecosystems; workers displaced without meaningful re-skilling pathways; energy use surging in areas already vulnerable to climate stress. 

What Companies Can Do Now

The path forward, according to O’Neill, is rooted in clarity, alignment, and iteration. Businesses don’t need to pivot overnight or rebuild their operations from scratch. They need to take stock of what they already do well, identify the SDG most aligned with their mission, and begin tracking meaningful, relevant metrics that reflect their contribution to a better future.

This can be as simple as adding one SDG-aligned KPI to a leadership dashboard or as complex as redesigning hiring practices to retain knowledge and community ties. What matters most is the intentionality behind the action.

For leaders struggling with how to begin, O’Neill offers practical guidance: don’t wait for perfect information. Move. Learn. Adapt. Align technology strategy with purpose — not in a silo, but as part of a larger ecosystem of human and planetary thriving. Because in the future of work, success will be defined by how wisely we integrated AI into the human systems that sustain us.

Advertisements

Subscribe to the Future of Work Newsletter





Source link

Continue Reading

Business

Fliggy’s Business Travel Arm Launches AI-Powered Solution

Published

on


Intelligent automation and analysis enhance compliance, cost savings, and employee experience

Hangzhou, China (ANTARA/PRNewswire)- AliBtrip, a designated platform specializing in business travel management under Fliggy of Alibaba Group (NYSE: BABA and HKEX: 9988), has introduced an integrated AI business travel solution. The innovative solution features two key modules: an employee travel agent for personalized planning, and a corporate management agent to streamline financial administration and compliance through data-driven decision-making support.

AliBtrip distinguishes itself from other Travel Management Companies (TMCs) with its unique AI applications and digital transformation management, leveraging Alibaba’s extensive ecosystem. Recent data indicates that AliBtrip serves over 20,000 industry-leading clients and more than one million growth companies, with over 20 million corporate employees booking business trips through the platform.

“The foundation of business travel services lies in trust between employees and companies,” said Zhuoran Zhuang, Vice President of Alibaba Group and CEO of Fliggy. “AliBtrip’s solution aims to transform AI capabilities into tangible benefits, reducing both visible and hidden costs while enhancing value for our clients.”

Addressing the challenges of corporate travel

Unlike consumer-focused solutions, the AI implementation for business travel emphasizes efficiency and compliance at every stage. AliBtrip’s AI solution, powered by multiple intelligent agents, tackles the complexities of corporate travel through a sophisticated division of labor among agents. It integrates long and short-term memory management and real-time deployment of the Model Context Protocol (MCP).

Powered by AliBtrip and Fliggy’s extensive data from the hotel and travel sectors, this AI solution draws from a real-time price database for flights, accommodations, transportation, and dining, ensuring practical and effective travel planning.

Optimizing management with intelligent analysis

For businesses, AliBtrip’s AI acts as an expert in administrative and financial management, offering real-time strategic analytics and action support. This capability significantly reduces the transactional workload related to analysis, communication, and compliance.

In addition to its pricing database, AliBtrip’s AI solution can customize exclusive datasets that reflect each corporate client’s travel standards and employee preferences, aligning with the company’s policies and values.

Features such as a strategy center and natural language interaction streamline corporate management, with intelligent cost control options presented in clear, quantitative indicators and intuitive examples for decision-makers, allowing them to make adjustments with a single click. The AI can also analyze historical travel data to identify potential issues proactively.

Enhancing employee satisfaction through streamlined booking

For employees, AliBtrip’s AI simplifies the booking process, alleviating the burden of comparing travel policies and booking transportation, accommodation, and car services separately. The employee travel agent generates comprehensive itineraries based on three key inputs: purpose, time, and destination, linking seamlessly to the travel request in the system. After verifying departure and arrival locations, it autonomously creates the itineraries that include tickets, hotels, and transportation, all while considering factors such as weather and check-in times.

Real-time travel assistance enhances the overall experience, with automatic reminders and recommendations integrated into the travel itineraries that comply with corporate standards. This significantly reduces risks associated with budget overruns or non-compliance.

The AI solution also uncovers cost-saving opportunities often-overlooked; for example, suggesting business class tickets with early departures that could avoid overnight stays or prioritizing hotels that align with employee preferences within budget constraints.

“Employees should be served, not restricted,” said Shenyang Shi, General Manager of AliBtrip, highlighting the philosophical shift underlying the innovative solution. “By leveraging advanced travel planning algorithms and combining intent recognition capabilities with comprehensive datasets and route optimization, the platform demonstrates how AI can reconcile cost management with employee satisfaction, creating value for both businesses and their traveling workforce.”

About Fliggy

Fliggy is a wholly-owned subsidiary of Alibaba Group (NYSE: BABA and HKEX: 9988 (HKD Counter) and 89988 (RMB Counter), and is one of the leading online travel platforms in China. Fliggy places a strong emphasis on innovation in its products and services, catering to the increasingly personalized and diversified needs of consumers in both China and overseas markets.

Leveraging Fliggy’s advantage as part of the Alibaba ecosystem, merchants can benefit from the vast user base within the Group. Fliggy also collaborates with partners through a full-service management format, helping more merchants, especially small and medium-sized ones, easily and efficiently share opportunities enabled by digitalization.

Fliggy’s long-term strategy is to promote the digital transformation of the tourism industry, using an open platform and mechanisms to help the industry make better use of digital business infrastructure for their operations.

Source: Fliggy

Reporter: PR Wire
Editor: PR Wire
Copyright © ANTARA 2025



Source link

Continue Reading

Trending