Advertising Tests and Company Dramas in the Artificial Intelligence Sector
In the 232nd episode of the Last Week in AI podcast, the week's highlights were discussed, including OpenAI beginning ad tests in ChatGPT, management changes at Thinking Machines, and new research models.
OpenAI Testing Ad Integration in ChatGPT
AI giant OpenAI has begun trialing advertisements in its popular chatbot, ChatGPT. This move is seen as part of the company's search for a new revenue model to cover its operational costs, which run into billions of dollars. Simultaneously, a system to estimate users' age has been implemented to strengthen the platform's security features.
Busy Days in Silicon Valley: Departures from Thinking Machines
AI startup Thinking Machines has experienced a series of high-profile employee departures that have made waves in the industry. It's reported that some of those who left the company have returned to OpenAI. This development is interpreted as an indicator of how intense competition and the talent war in the AI industry can rapidly change organizational dynamics. Similar activity was also observed in merger and funding news within the health tech sector.
Research and Technological Developments
The podcast also introduced a new research model called 'STEM: Scaling Transformers with Embedding Modules'. This model aims to increase the efficiency and scalability of large language models. On the other hand, developments on the hardware side, such as Intel's advances in visual processing technologies, are also critical for the future of AI.
Regulation and Security Agenda
The US Senate passed the 'Defiance Act', a law granting victims the right to sue over explicit content generated by artificial intelligence. This shows that legal regulations concerning AI ethics and content moderation are gaining momentum. Meanwhile, the company Anthropic updated the 'constitution' that defines the ethical rules for its AI assistant, Claude. User-experience-focused updates are also being seen on different platforms, such as the 'break' reminder coming to Google's Gemini.
Sectoral Activity and Market Dynamics
China's Zhipu AI took a significant step in breaking dependence on US chip supply by training its advanced models entirely on domestic Huawei hardware. Meanwhile, Elon Musk's xAI activated the world's first Gigawatt-level AI supercluster to compete with OpenAI and Anthropic. On the investment front, it's reported that venture capital firm Sequoia plans to invest in Anthropic, breaking its taboo of investing in rival firms.
Artists have launched the 'Stealing Isn't Innovation' campaign to protest big tech companies using their work without permission to train AI models. This situation has reignited global debates on intellectual property rights, alongside the proliferation of creating images with AI.
