LEVITTOWN, N.Y., Sept. 27, 2023 (GLOBE NEWSWIRE) -- Cogito Tech, a trusted leader in data labeling for AI development, is taking the lead in promoting a more ethical approach to artificial intelligence (AI), machine learning (ML) development and employment practices. Today, Cogito launched an appeal to AI vendors worldwide by unveiling a “Nutrition Facts” style framework for AI training data called DataSum. Cogito Tech challenges the industry to pull back the curtain on the human effort, fueling AI’s advancements with a call to action to adopt more humane labor practices and ethical data curation strategies.
Cogito and DataSum
Cogito aims to provide the highest level of transparency and ethical considerations in the data lifecycle. With DataSum, the company encourages stakeholders to leverage this framework to make informed decisions on dataset selection, workforce conditions, and technology stack for data labeling. The newly introduced DataSum aims to make these commitments transparent and measurable, serving as an AI supply chain management component and benchmark for the entire industry. As a result, this level of detail allows users to make more informed and ethical decisions about the AI they employ.
“As we continue to innovate in AI, the need for a trustworthy and comprehensive Supply Chain Management framework becomes ever more critical,” says Rohan Agrawal, CEO of Cogito Tech. “DataSum is our answer to this urgent need. The potential to revolutionize consumer interactions with AI technologies across various domains, from personalized medical treatments and comprehensive financial planning to legal insights, is paramount.
“By harnessing the collective expertise of our industry, we have the opportunity to establish a new standard—one that enhances and safeguards the AI experience for consumers worldwide,” continues Agrawal.
Labor Abuses in Data Labeling
While AI adoption and innovation are skyrocketing, much of the labor force behind AI is working for low pay and unfavorable conditions. For instance, human data labelers in Kenya, the Philippines, India, and other regions are often required to work for less than $2 per hour.
“AI doesn’t work in a vacuum. AI relies on human-crafted data for proper context and accuracy, but the role of data labelers is undervalued,” says Matthew McMullen, Senior Vice President of Cogito Tech. “Without skilled human expertise to shape and refine the data, the AI algorithms will be increasingly inaccurate. It’s time the AI industry recognized the contribution of data labelers by improving working conditions and paying a living wage.”
For instance, RLHF (Reinforcement Learning with Human Feedback) is essential to machine learning. However, these gig workers who perform such critical data labeling assignments earn less. A recent report on online outsourcing from the University of Oxford’s Internet Institute reported that none of the organizations surveyed met the minimum standard for fair labor principles (fair pay, fair conditions, fair contracts, fair management, and fair representation). The report also showed that workers earned an average of $2.15 per hour, and nearly 27% of their work time consisted of unpaid tasks.
Workforce Contracts and Career Mobility
Additionally, laborers advancing AI technologies find themselves often caught in strict contractual agreements that workers are compelled to sign. These contracts often grant the data vendor exclusive ownership of a worker's professional identity.
“These workers are not just cogs in the wheel but vital contributors. The loss of control these workers experience over contractual agreements leaves them unable to negotiate for better terms and seek higher-paying opportunities. The existing system often marginalizes them and hinders their career mobility. A call for fair wages based on experience, plus genuine opportunities for skill development and career progression, must be instituted - this is what we are working toward at Cogito,” says McMullen.
International guidelines and emerging standards such as the International Labor Organization (ILO) and the ILO Tripartite Declaration have made strides in advocating for worker rights and fair wages. These guidelines call for businesses, including data vendors, to implement ethical practices that ensure workers are compensated fairly, but there are still critical challenges these data workers must endure.
The Need for Ethical Data Sourcing
In addition to ethical labor practices, Cogito calls for data vendors to adopt higher standards in sourcing datasets. Protecting intellectual property is a growing concern with advances in AI, and Cogito is calling on AI vendors to commit to ethically sourcing data across all sectors.
Ethically sourced data includes avoiding proprietary data and copyrighted material as part of data labeling. In addition, using data from other systems can introduce biases, highlighting the need to identify data origins and the potential implications for machine-generated data.
Incorporating an “AI supply chain management” approach, Cogito with DataSum insists on tracking the ethical implications of data from its source through its various transformations.
“AI has breathtaking capabilities, but while AI can imitate human discernment, thinking, and processes, its performance hinges on the quality of the data it receives,” says McMullen. “Large Language Models (LLMs) may generate seemingly genuine information, but this can crumble when faced with deceptive inputs, demonstrating AI's lack of true cognition and emphasizing the importance of human intervention in its training.”
“If AI is going to advance humanity, then we need to balance AI’s potential with ethics,” continues McMullen, “We need to change the narrative and acknowledge that AI is more than cutting-edge algorithms but the result of human dedication and ingenuity. The AI industry must adopt a code of ethics that ensures workers are properly compensated for their expertise, and datasets are responsibly sourced.”
McMullen stresses that enterprises that engage AI vendors play a critical role in championing labor standards, ethical sourcing, and sustainability.
About Cogito Tech:
Since 2011, Cogito Tech has become a leading AI training data company, offering human-in-the-loop workforce solutions comprising Computer Vision, Natural Language Processing, Content Moderation, data and document processing. Cogito’s mission is to embrace the power of human ingenuity and technology to create 360* value for AI and Business Initiatives. The company’s vision is to support the development of game-changing AI and Technology Applications by providing cutting-edge workforce solutions to solve everyday business needs. For more information, visit www.cogitotech.com.
Contact:
Michele Nachum
Firecracker PR
michele@firecrackerpr.com
Photos accompanying this announcement are available at
https://www.globenewswire.com/NewsRoom/AttachmentNg/3835e7fe-4710-4c3c-8c15-59330e05ba9e
https://www.globenewswire.com/NewsRoom/AttachmentNg/52854600-3963-4661-80af-2e8aac0aa394
