Staying updated in the fast-paced world of AI can be challenging. To help you stay informed, here is a roundup of recent stories in the field of machine learning and notable research that you may have missed.
This week in AI, the focus is on labeling and annotation startups, such as Scale AI. Scale AI is reportedly in discussions to raise new funds at a valuation of $13 billion. While labeling and annotation platforms may not grab headlines like cutting-edge generative AI models, they play a crucial role in the development of AI models.
Data annotation is essential for training AI models, as labels help the models interpret and understand data during the training process. Without accurate and high-quality labels, the performance and reliability of AI models can be compromised.
However, the treatment of data annotators in many startups is often subpar, with low wages and poor working conditions prevalent in the industry. Some annotators, particularly in third-world countries, are exposed to disturbing content without adequate support.
An investigation into Scale AI revealed harsh working conditions for annotators, with long hours and low pay being common. The lack of standard regulations for ethical labeling work poses challenges for the industry.
Policymaking may be the key to addressing these issues, as self-regulation by platforms has proven ineffective. Here are some other recent AI stories worth noting:
- OpenAI builds a voice cloner: OpenAI introduces Voice Engine, a tool allowing users to clone voices from short recordings.
- Amazon invests in Anthropic: Amazon increases its investment in AI power Anthropic by $2.75 billion.
- Google.org launches an accelerator: Google.org launches a program to fund nonprofits developing generative AI tech.
- New model architecture: AI21 Labs releases Jamba, a generative AI model using state space models.
- Databricks releases DBRX: Databricks unveils a new generative AI model for various AI benchmarks.
- Uber Eats and UK AI regulation: Explore the challenges of AI bias in platforms like Uber Eats under UK regulations.
- EU election security guidance: The EU publishes draft guidelines on election security to combat disinformation.
- Grok gets upgraded: X’s Grok chatbot receives an upgraded model, Grok-1.5, available to all Premium subscribers.
- Adobe expands Firefly: Adobe introduces Firefly Services, offering new APIs and tools for creatives.
More machine learnings
Recent advancements in AI include SEEDS, a system for scalable weather forecasting, and Fujitsu’s application of AI to underwater imagery. Linear functions have been found to play a significant role in large language models, impacting their recall mechanisms.
Improving character interactions through AI, as demonstrated by Disney Research, and the importance of addressing bias in AI and search ethics are crucial areas of focus in the evolving landscape of AI and technology.
