This week in artificial intelligence (AI) news, experts weigh in on Elon Musk’s prediction that AI will soon outsmart humans, companies are building custom AI chips to cut costs, analysts say the demand for quality data to feed AI programs will soon surpass supply and the mental health sector is seeing a boom in AI-powered apps.
Is AI going to outsmart us? Elon Musk recently said artificial intelligence (AI) could surpass the intelligence of the brightest humans as early as next year or by 2026, sparking a robust debate among scholars, technologists and ethicists.
In an interview on X, the Tesla CEO discussed the quickening pace of AI development aimed at achieving and surpassing human cognitive capabilities. Experts are now examining the likelihood of Musk’s forecast and the significant questions it poses about the nature of intelligence, ethical limits and the future dynamics between humans and machines.
“Elon is correct,” said Yigit Ihlamur, an AI researcher and founder of Vela Partners, an AI-focused investment firm, in a conversation with PYMNTS. “AI has already surpassed human intelligence in certain areas and will do so in more areas — though not all.”
AI chips are all the rage these days. Major tech companies are actively developing their own custom chips to enhance the efficiency and reduce the costs of artificial intelligence (AI) operations.
Meta recently introduced its latest custom chips designed to boost its AI capabilities and reduce reliance on third-party suppliers like Nvidia. This move aligns with Intel’s recent unveiling of an advanced AI “accelerator,” and mirrors similar initiatives by Google to produce AI chips internally. Experts believe these chips could enhance commercial AI applications.
“Custom chips lower the threshold for businesses to train AI models tailored to specific customers and tasks, moving beyond merely using APIs from large language model providers, especially in specialized and high-security scenarios,” said Amrit Jassal, co-founder and CTO of Egnyte, a company that develops AI-driven software for businesses, in a discussion with PYMNTS.
Additionally, custom chips may significantly cut AI-related costs for companies. Itrex, a software development firm, said incorporating generative AI into business operations currently ranges from a few hundred to several hundred thousand dollars a month for bespoke solutions using fine-tuned open-source models. Last year, Nvidia CEO Jensen Huang highlighted the potential cost savings from custom AI chips.
Data is the lifeblood of AI, and it’s running dry. Industry analysts warn that the escalating demand for high-quality data, crucial for powering AI conversational tools like OpenAI’s ChatGPT, could soon exceed supply and potentially hinder AI advancement.
The increasing dependence on extensive datasets poses a challenge for AI development. Such data is vital for refining models like ChatGPT, but the looming scarcity of data is causing concern among tech professionals.
The dearth of training data for AI results from the need for large quantities of diverse, high-quality and accurately labeled data that mirrors real-world situations the models will face. Collecting this data is often labor-intensive, requiring manual annotations by experts, sourcing from various locations, and meticulous curation to maintain quality and remove biases.
Moreover, AI companies encounter significant copyright hurdles in gathering training data, necessitating careful adherence to legal requirements, obtaining permissions and filtering content.
“Humanity cannot replenish this resource as quickly as LLM companies are depleting it,” said Jignesh Patel, a computer science professor at Carnegie Mellon University and co-founder of DataChat, a generative AI analytics platform. “However, specialized LLMs rely less on publicly available data. For example, an LLM designed to automate a financial risk review process might be specific to a single bank or investment firm, often with little to no public documentation available.”
The urgency of securing training data is underscored by lawsuits from authors and publishers against AI firms, accusing them of illegally using their content to develop these technologies.
Your next therapist could be a chatbot. The mental health field is experiencing a surge of interest in AI-powered apps, promising rapid and accessible care in a sector with growing demand.
These AI applications are stirring debates among professionals about their potential to fill care gaps and concerns regarding their effectiveness and ethical implications. The key challenge is determining if AI can truly augment the human element in mental health treatment.
Derek Du Chesne, CEO of Better U and creator of its mental wellness app, told PYMNTS AI can have a role in personalizing care. “AI is invaluable in mental health apps because it can tailor support and interventions for each user based on their data, improving engagement and outcomes,” he said. “AI algorithms can detect patterns in behavior and mood over time, allowing for early intervention and proactive care.”
Du Chesne also noted that AI provides round-the-clock support, which is essential when human therapists are unavailable.
This trend is part of a broader move toward using AI in customer service. As reported by PYMNTS, consumers frequently interact with AI technologies in various applications, from preventing credit card fraud to handling returns via chatbots. Meeranda, a company based in Toronto, is preparing to launch a visual AI designed to emulate real-time human interactions, moving beyond typical chatbot functions and not directly competing with platforms like ChatGPT.
We’re always on the lookout for opportunities to partner with innovators and disruptors.
Learn More