AI Roundup — May 17, 2026
ArXiv Moves to Ban Authors Who Rely Entirely on AI for Research Papers
Scientific preprint repository ArXiv has announced stricter enforcement measures targeting the careless use of large language models in academic submissions, according to TechCrunch. Under the updated policy, authors found to have delegated the entirety of their written work to an AI system face a one-year ban from submitting to the platform.
ArXiv's stance draws a distinction between using AI as an assistive tool versus using it as a wholesale replacement for original scholarly writing. The repository has not prohibited AI assistance outright, but is moving to hold authors accountable for the integrity and authenticity of their submissions. The move reflects a broader effort within academic publishing to establish clearer standards around AI-generated content in research contexts.
OpenAI's Greg Brockman Steps Into Product Strategy Role; ChatGPT and Codex Merger Reported
OpenAI co-founder Greg Brockman has taken on responsibility for the company's product strategy, TechCrunch reports. The leadership shift coincides with reports that OpenAI is planning to combine two of its flagship offerings — ChatGPT, its general-purpose conversational AI, and Codex, its programming-focused product — into a unified experience.
Brockman is one of OpenAI's original founders and has held various technical and operational roles at the company over the years. The reported consolidation of ChatGPT and Codex would mark a notable product direction change, potentially streamlining how developers and general users access OpenAI's capabilities through a single interface. OpenAI has not issued a detailed public announcement on the timeline or scope of the integration as of this writing.
Sony Clarifies How Its AI Camera Assistant Works on the Xperia 1 XIII
Following public attention around a demonstration post, Sony has issued clarifications on the functionality of the AI Camera Assistant featured in its Xperia 1 XIII smartphone, The Verge reports. Sony states that the feature does not alter or edit photographs after they are taken. Instead, it analyzes scene conditions — including lighting, depth, and subject — and surfaces four compositional or settings suggestions to the user before the shot is captured.
The clarification appears intended to address concerns that the tool was modifying images in ways that could be considered non-transparent. Sony's explanation positions the AI Camera Assistant as a real-time advisory layer rather than a post-processing filter, keeping the final creative decision with the photographer.
These developments span research integrity policy, enterprise product consolidation, and consumer hardware AI features — reflecting the continued breadth of AI's integration across industries this week.