This week’s AI news roundup covers a broad range of stories, from a mysterious chatbot’s brief appearance, AI safety efforts, and digital music generation, to the potential for AI-powered search platforms and autonomous killer robots.
The unnamed chatbot, dubbed ‘gpt2-chatbot,’ surfaced on the LMSYS Chatbot Arena, impressing users with its performance that seemed to surpass that of GPT-4. However, it was quickly removed, causing curiosity and speculation around who created it and whether it could be a precursor to GPT-5.
An examination of OpenAI’s SSL logs revealed a hint at a new search platform, sparking dialogue around the potential for AI to take on Google. Further, music AI generators are changing the musical landscape, as tools like Udio and Suno enable users to generate and upload AI-created music to platforms like Spotify.
When it comes to AI safety, there seems to be a lack of consensus on how to protect humanity from AI-fueled threats. The first AI Safety Summit at Bletchley Park led to a global commitment to AI safety, though skepticism arose when several influential figures stated they would not attend the follow-up meeting.
At a two-day event in Austria, representatives from 143 countries hashed out the pros and cons of autonomous weapons systems. The discussion fueled concerns about the potential dangers of advanced AI technology, leading to calls for regulation of the burgeoning field.
Adding to the AI saga, a Baltimore school principal was the victim of a calculated AI attack. A former gym teacher manipulated AI technology to mimic the principal’s voice and make offensive comments, resulting in the principal’s temporary job loss while investigations took place.
Legal battles continue to unfold around OpenAI’s use of internet data for training its GPT models, as the company negotiates a deal with The Financial Times for training material. As AI continues to evolve and proliferate, the discussions surrounding data usage and ‘fair use’ continue unabated.
On the global stage, Soheil Zabihi from TokenScope offered insight into how AI and blockchain technology could revolutionize online transactions. Other notable updates include Google DeepMind beating transformer models, Apple releasing a more accurate language model known as OpenELM, and China unveiling Vidu, a powerful text-to-video generator.
This eventful week in AI news culminates in questions about the future of AI, from the potential for AI-run search platforms to AI-generated music, and the prospects and perils of autonomous killer robots.
