Generative AI needs to be more inclusive as it evolves, according to OpenAI CEO

Generative artificial intelligence (AI) tools like ChatGPT will need to be more inclusive as they continue to develop, so that they can be accepted by a wider global population.

According to Sam Altman, CEO of OpenAI, the brainchild behind ChatGPT, the emerging technology should also be developed with public consultation, with humans remaining in control. This was necessary to mitigate the potential risks or pitfalls associated with AI adoption, Altman said during an interactive session in Singapore, which was part of the CEO’s month-long global tour that also included regional markets South Korea, Indonesia, Japan were involved. , and Australia.

Too: Generative AI: Don’t call it ‘artist’, say scholars in Science journal

Organized at the Singapore Management University and organized by AI Singapore and Infocomm Media Development Authority, the interactive session was attended by students, developers and industry players.

Elaborating on how risks might be managed as AI evolves, Altman said it was important to let the public learn about and experience new developments as they came along. This will ensure that any potential damage is exposed and addressed before its effects become widespread.

This would be more cost effective than building and testing a piece of technology behind closed doors, and releasing it to the public on the assumption that all potential risks were identified and plugged.

Too: 5 Ways to Explore the Use of Generative AI at Work

“You can’t learn everything in a lab,” he said. No matter how well a product is tested for the purpose of reducing harm, someone will think of ways to exploit it that its creators never thought of. He said this was true with any new technology.

“We believe that iterative deployment is the only way to do this,” he said, adding that introducing new releases gradually also enables society to incorporate feedback on how AI evolves and how AI should be improved. Is.

Pointing to the rapid rise of ChatGPT as how it started a worldwide conversation about the risks of AI, Altman said such discussions were important and would not have come to the fore if the generative AI platform had not been deployed .

Too: Leadership Alert: The Dust Will Never Settle and Generative AI Can Help

As the allure of AI grows and the interest of nations grows, it will also be important to address challenges related to bias and data localisation.

Altman said that for OpenAI, this means figuring out how to train its generative AI platform on datasets that are “as diverse as possible.” These need to cut across multiple cultures, languages ​​and values.

While it is impossible to build a system that everyone collectively agrees is fair, it is important for AI models to learn as much as possible and allow users to retain control. He underlined the need for humans to have a tight loop and not be too trusting, while giving AI systems full control over decision-making.

he also reiterated call for governments As for working out the rules around AI, adding that until different jurisdictions work out what they plan to do, it will be difficult to determine which framework will work best.

He suggested that what could emerge was a mix of scenarios where some countries could come together to establish a common legal framework, while organizations could establish industry standards and best practices to adopt. Citizens groups can also suggest codes of practice.

Too: GitHub developer survey says 92% of programmers are using AI tools

In the meantime, OpenAI will continue to work on ChatGPT releases that are dramatically better than previous ones, Altman said, noting that more videos and images will be added to the platform’s learning base.

He said the company aims to enhance ChatGPT’s role as a tool for humans and their ability to perform tasks.

Other data systems matter too

However, organizations looking to adopt generative AI must remember not to do so in isolation from other systems.

For example, chatbots, which are powered by generative AI models, should be integrated with a company’s customer relationship management (CRM), said Jun Yang, Google Cloud’s VP of cloud AI and industry solutions.

Yang said during a virtual media briefing to provide an update about the US vendor’s generative AI plans that customers often visit the company’s website with product questions or questions related to their accounts. This means that the most recent and relevant information needs to be pulled from CRM systems and product databases.

The data used to train large language models is unlikely to be up to date to power generative AI systems and chatbots may push out-of-date information to customers.

Too: Is humanity really over? Consider AI’s Achilles Heel

She noted that Google’s customers have opted to take a more cautious approach with generative AI, being the first to remind employees that not all responses generated will be accurate. He added that other customers, especially startups, have chosen to tap generative AI to push the boundaries and transform their industry.

Asked if there was a lack of localization, especially for businesses in Asia, since AI development was mostly led by US players and models trained on data outside Asia, Yang acknowledged that most innovation came from the US. Were. However, he noted that significant academic research in AI is being conducted in Europe and Asia.

Google has also trained its models to support multiple languages. For example, its speech-to-text platform Chirp is currently capable of handling more than 100 languages, he said. Chirp’s foundation model has been integrated into Google’s machine learning training and development platform, Vertex AI, which was recently expanded to include generative AI support. Foundation models are accessible through APIs (Application Programming Interfaces).

Too: According to Research, Generative AI Can Save Marketing Professionals 5 Hours Per Week

Yang said Google has been open about sharing its technologies and research results, which it is happy for others to leverage and extend to drive their own AI innovation. He added that US vendors are also ready to cooperate with partners.

Elsewhere, in China, Alibaba Cloud introduced a partnership program in April to build generative AI models that are customized for companies across verticals including finance and petrochemicals. The Chinese cloud vendor hopes to accelerate the development of applications that are powered by its larger language model, Tongyi Qianwen, that will integrate with all of Alibaba’s business applications, including e-commerce, search, navigation and intelligent voice assistance.