High-quality code documentation is crucial for software development especially in the era of AI. However, generating it automatically using Large Language Models (LLMs) remains challenging, as existing approaches often produce incomplete, unhelpful, or factually incorrect outputs. We introduce DocAgent, a novel multi-agent collaborative system using topological code processing for incremental context building. Specialized agents (Reader, Searcher, Writer, Verifier, Orchestrator) then collaboratively generate documentation. We also propose a multi-faceted evaluation framework assessing Completeness, Helpfulness, and Truthfulness. Comprehensive experiments show DocAgent significantly outperforms baselines consistently. Our ablation study confirms the vital role of the topological processing order. DocAgent offers a robust approach for reliable code documentation generation in complex and proprietary repositories.

We introduce a novel multi-agent collaboration framework designed to enhance the accuracy and robustness of text classification models. Leveraging BERT as the primary classifier, our framework dynamically escalates low-confidence predictions to a specialized multi-agent system comprising Lexical, Contextual, Logic, Consensus, and Explainability agents. This collaborative approach allows for comprehensive analysis and consensus-driven decision-making, significantly improving classification performance across diverse text classification tasks. Empirical evaluations on benchmark datasets demonstrate that our framework achieves a 5.5% increase in accuracy compared to standard BERT-based classifiers, underscoring its effectiveness and academic novelty in advancing multi-agent systems within natural language processing.

While the biases of language models in production are extensively documented, the biases of their guardrails have been neglected. This paper studies how contextual information about the user influences the likelihood of an LLM to refuse to execute a request. By generating user biographies that offer ideological and demographic information, we find a number of biases in guardrail sensitivity on GPT-3.5. Younger, female, and Asian-American personas are more likely to trigger a refusal guardrail when requesting censored or illegal information. Guardrails are also sycophantic, refusing to comply with requests for a political position the user is likely to disagree with. We find that certain identity groups and seemingly innocuous information, e.g., sports fandom, can elicit changes in guardrail sensitivity similar to direct statements of political ideology. For each demographic category and even for American football team fandom, we find that ChatGPT appears to infer a likely political ideology and modify guardrail behavior accordingly.

The CEO of OpenAI, Sam Altman, has spent the last month touring world capitals where, at talks to sold-out crowds and in meetings with heads of governments, he has repeatedly spoken of the need for global AI regulation. But behind the scenes, OpenAI has lobbied for significant elements of the most comprehensive AI legislation in the world--the E.U.'s AI Act--to be watered down in ways that would reduce the regulatory burden on the company, according to documents about OpenAI's engagement with E.U. officials obtained by TIME from the European Commission via freedom of information requests. In several cases, OpenAI proposed amendments that were later made to the final text of the E.U. law--which was approved by the European Parliament on June 14, and will now proceed to a final round of negotiations before being finalized as soon as January. In 2022, OpenAI repeatedly argued to European officials that the forthcoming AI Act should not consider its general purpose AI systems--including GPT-3, the precursor to ChatGPT, and the image generator Dall-E 2--to be "high risk," a designation that would subject them to stringent legal requirements including transparency, traceability, and human oversight. That argument brought OpenAI in line with Microsoft, which has invested $13 billion into the AI lab, and Google, both of which have previously lobbied E.U. officials in favor of loosening the Act's regulatory burden on large AI providers.

An animal-welfare organization said it plans to ask a U.S. government agency on Thursday to investigate Elon Musk's brain-implant company Neuralink over records it said show potentially illegal movement of hazardous pathogens. The Physicians Committee of Responsible Medicine (PCRM) said in a letter to the U.S. Department of Transportation, which was shared with Reuters, that it has obtained emails and other documents that suggest unsafe packaging and movement of implants removed from the brains of monkeys. These implants may have carried infectious diseases in violation of federal law, PCRM said. The letter said records that the group obtained showed instances of pathogens, such as antibiotic-resistant staphylococcus and herpes B virus, that may have been transported without proper containment measures. PCRM's letter adds to the scrutiny facing Neuralink, which is developing a brain implant it hopes will help paralyzed people walk again and cure other neurological ailments.

The European Parliament's co-rapporteurs of the AI Act have proposed expanding the European Commission's revision powers to extend the list of high-risk systems and prohibited practices at a later stage. The eighth batch of compromise amendments on the proposed Artificial Intelligence regulation was shared by leading lawmakers Dragoș Tudorache and Brando Benifei with the representatives of the other political groups last Friday (21 October). The AI Act is a flagship legislative proposal to regulate the development, deployment and use of artificial intelligence. The co-rapporteurs are currently trying to reach a common position with the other political groups, pitching several significant changes to the original wording. The most relevant amendment would significantly extend the Commission's revision powers after the legislation becomes effective.

The police are failing to consult the public about their growing use of technologies including artificially-intelligence facial recognition and automated decision systems (ADS), a charity has warned. South Wales Police is the only police force in the UK known to be using AI in its policing to have confirmed it consulted with its local communities about its use, according to a report from The Royal Society for the encouragement of Arts, Manufactures and Commerce (RSA). A Freedom of Information request returned in March found that London's Metropolitan Police Force, which began using live facial recognition tech in February following years of trials, had no record of consulting the public, despite suggesting that this would take place alongside deployment. The Met's software is deployed through signposted cameras focused on small areas to scan the faces of passers-by in areas the force believes are more likely to contain those wanted for serious and violent offences. The RSA sent requests to 45 territorial police forces, receiving confirmation that eight were using or trialling AI or ADS for policing decisions, including Durham Constabulary, Surrey Police and West Yorkshire Police.

Over half of NHS Trusts (52%) are already deploying artificial intelligence (AI) technologies, with 20% of them using them for clinical care, and 16% using them for clinical diagnosis, a Freedom of Information request has revealed. The data was obtained by a Freedom of Information request from cloud data services provider NetApp, with 61 Trusts responding, after the UK Government announced its investment in AI in the National Health Service (NHS). The request asked Trusts around their current and future use of AI-related technologies to deliver health services, and how mature their data infrastructures are in delivering successful AI projects. George Kurian, NetApp CEO and president, said: "Artificial intelligence has limitless potential in healthcare services and it's encouraging to see the technology being used in half of NHS Trusts. As healthcare moves towards preventative treatment and personalised medicines, artificial intelligence leaders in the NHS have a complex challenge to break through cultural and organisational barriers when it comes to providing healthcare professionals the access to data they require. "Progress is being made and the further deployment of AI-powered technologies – such as speech recognition and machine learning – will alleviate pressure on staff, accelerate innovation and reduce costs.

Manulife Financial Corp. said Thursday it will cut about 700 jobs as it becomes the latest financial services company to streamline and digitize customer service operations. The cuts will largely target customer service positions that are no longer necessary as the company automates customer transactions, said Manulife Canada CEO Michael Doughty. "Our industry, including us, are still doing too many things the old way: processing paperwork, accepting mail, answering telephone calls on information requests that clients should be able to access on their own." "This is a pretty bold step in transforming ourselves to become a digital, customer-centric organization," said Doughty. The job cuts will come through voluntary exit programs and natural attrition over the next 18 months, the financial services company said.

When a citizen dials 311, it has been the longstanding preference by mayors to have a city employee on the other end of the line to deliver the ever-valuable personal touch. But when efficiency is the priority, are we really best served by having city employees at 311 call centers act primarily as switchboard operators, sifting through online scripts or, worse, binders or spreadsheets and responding to information requests with specifically coded responses? More often than not, citizens will call 311 with an information request -- to determine their trash pick-up day, the hours the public pool is open or another simple ask. The North Carolina Innovation Center is now using chatbots for its internal IT help desk hotline, where between 80 and 90 percent of calls are for help changing a password. Still in its experimental stage, North Carolina uses the bots to free up help center operators to handle more challenging and complex concerns.