ChatGPT Buzz Kill
The buzz around ChatGPT-like chatbots is slowly diminishing as tech giants such as OpenAI and Google emphasize on AI safety rather than releasing improved models. Sam Altman, the CEO of OpenAI mentioned that GPT-5 won’t be coming out any time soon, while Google mentioned that user feedback and safety measures will help them improve their rival chatbot, Bard.
Could the tech giants’ seemingly slowing down on releasing advanced capability chatbots be an indication of AI development hitting a plateau? How is the industry prioritizing safety over functionality?
Safety first
With problems such as hallucinations and non-factual data, chatbots are not yet perfect, so companies must take into account security and safety issues when developing these chatbots. Google released Bard with a “Google it” button to address hallucinations, as well as other safety filters to filter out hate speech and bias. Sundar Pichai of Google recently explained how AI systems must be aligned with human values, and that achieving this requires input from more than just engineers, such as ethicists, philosophers and social scientists. Google’s rushed approach to the AI chatbot space resulted in Bard giving an incorrect answer at the Paris AI event, leading to a $100 billion drop in their market value.
OpenAI’s ChatGPT, while showing some promising features, has been met with issues. An open letter was signed by over 13,500 people including Elon Musk and Steve Wozniak, to pause the release of models beyond GPT-4 to account for the “profound risks to society and humanity”. OpenAI has made multiple statements to address AI safety in the wake of this letter. Sam Altman highlighted that the company will be moving forward with caution and increased rigour when it comes to safety.
Less Noise, More Progress
Meta, a company that has been slow to the chase with LLMs, announced LLaMA in February. The foundation language model ranges from 7B to 65B parameters. They have forayed into the LLM space in the past with products such as BlenderBot and Galactica, however, the poor performance with hallucinations and racist responses led to failure.
Microsoft has been the biggest winner with ChatGPT, powering their search engine Bing, as well as integrating on their cloud computing partner and enterprises. The announcement that Samsung will replace Google with Microsoft’s Bing Chat as the default search on their devices led to Alphabet’s stocks crashing and a loss of $57 billion.
Chatbots and LLMs are being used across various industries, however, going by the statements and interviews of OpenAI and Google, the chatbot race will see a certain slowdown. With the focus being on improving safety features, new models won’t be released in the immediate future. Even though the chatbot buzz is subsiding, others such as Meta are handling it smartly while OpenAI and the rest are being cautious with the developments.
The Person:
Sundar Pichai is CEO of Google and Alphabet Inc. Born in 1972 in Chennai, India, Mr. Pichai attended the Indian Institute of Technology Kharagpur and Stanford University before entering the tech industry. He joined Google in 2004, taking on product management roles before being promoted to vice president, working on Google Chrome and App Engine. Pichai was appointed CEO of Google and Alphabet Inc in December 2019, and has been credited with spearheading Google’s AI-first approach which has put the company several years ahead of others. Pichai is well known for his responsible approach to technology, and has focussed a large portion of his work on AI safety and security.
The Company:
OpenAI is a research lab based in San Francisco, California, that focuses on AI-related technologies and services. Founded by Sam Altman, Greg Brockman and Ilya Sutskever in 2015, OpenAI was created as a non-profit research group that aims to make sure AI benefits humanity. The company’s primary goal is to maximize new improvements in artificial general intelligence by making sure AI technologies are secure, safe, and trustworthy. OpenAI has developed several projects such as ChatGPT and GPT-4 which have become popular in the industry for their capability for natural language processing and generation. OpenAI also recently collaborated on a project with Microsoft which aims to develop AI applications using the Microsoft Azure platform.
