New studies show the limitations of relying on AI tools for health answers. Platforms must tighten safeguards and educate users on smart prompting.

The news: Meta and Midjourney formed a partnership to bring more image-generation tools to Facebook and Instagram. Meta is licensing Midjourney’s “aesthetic technology” for users and brands, Meta chief AI officer Alexandr Wang posted on Threads. He implied that the agreement may go past licensing and involve collaboration with Meta’s research teams to integrate Midjourney into future models and products. Our take: Brands should experiment with Midjourney to streamline content creation for Meta campaigns. However, they should also monitor outputs carefully for quality and copyright issues, especially considering Midjourney has faced allegations of IP misuse. Fast creation is only an advantage if it doesn’t trigger legal or reputational backlash.

The news: OpenAI is bringing its newest models to Amazon Web Services (AWS) for the first time, marking a major milestone in the ongoing battle for AI cloud dominance. Its open-weight models, gpt-oss-120b and gpt-oss-20b—available via Amazon Bedrock and SageMaker AI platforms—are able to handle complicated text-based operations and integrate into cloud-based systems. Our take: OpenAI’s models are getting easier to access, meaning lower costs and fewer technical hurdles to trying powerful AI tools. AWS customers should start testing oss-120b and oss-20b for things like generating subject lines, social copy, and campaign variations and explore ways to fine-tune the models with company data.

The news: Meta is in talks to invest upwards of $10 billion in Scale AI, a data labeling startup. The deal would be Meta’s biggest ever external AI investment and could help it position its Llama large language model (LLM) as an industry standard, per Bloomberg. Scale AI has already partnered with Meta to develop Defense Llama, an LLM designed for military use that’s built on Llama 3, and also works with Meta competitors like Microsoft and OpenAI. Our take: Meta’s massive investment could draw antitrust scrutiny in an era of acqui-hires. The outcome of active probes in Big Tech partnerships could influence regulatory action, especially if this investment contains any exclusivity that limits model training resources for other companies.

Separating research from product lets Meta build faster, think longer term, and prove it’s serious about becoming more than a social media company.

Meta’s CTO claims nimbleness trumps incumbency in AI. However, Meta’s AI future still depends on the health of its core ad business.

Its new tool helps advertisers dodge harmful content without overblocking, giving brands clearer control over where their ads show up.

A scrappy Chinese startup disrupts the AI hierarchy with a lean, open-source model, challenging Big Tech’s costly infrastructure and rewriting the rules of AI competition.

The partnership, hosted by AWS, could solidify Anthropic’s reputation as a trusted AI provider for high-risk government applications.

The definition of open-source AI now requires transparent training data, giving Meta incentive to clarify what information it feeds Llama.

By integrating Reuters content, Meta aims to enhance its AI’s credibility and drive ad revenue while carefully reentering the news domain.

From voice-activated AI to cutting-edge wearables, Meta's latest updates signal a push to dominate emerging tech markets.

Generative AI is the biggest technology introduction since at least the smartphone. This report looks at five ways the tool is transforming the marketing landscape.

With Bedrock, Amazon courts an array of AI startups: The platform offers a smorgasbord of generative AI tools for enterprise. But moves by Databricks and Meta could take model training in-house.