Content Fans

Category

AI News & Trends

1402 articles • Page 63 of 94

OpenAI opens ChatGPT ads to all US businesses, cuts $50,000 minimum

OpenAI opens ChatGPT ads to all US businesses, cuts $50,000 minimum

OpenAI has made its ChatGPT Ads Manager available to all U.S. businesses, removing the previous $50,000 minimum and agency-only restriction. Now, any verified advertiser can run campaigns with smaller budgets, and ads are priced per click, with suggested bids of $3-5. Early data from small business tests is limited, but initial results suggest low click numbers and possibly high conversion rates. Experts say more data may become available soon, and some recommend using organic methods to appear in ChatGPT while paid results are still being tested. The ad platform may be best for niche or high-value products until more performance data is available.

Hackers poison AI assistant descriptions, exfiltrate data

Hackers poison AI assistant descriptions, exfiltrate data

Hackers may be using a new tactic where they hide harmful instructions in AI assistant app descriptions to steal data, and this method appears to work on many popular tools. Reports suggest that these attacks succeed in about 60-72 percent of tested AI assistants, showing the risk might be widespread. Other poisoning techniques described include hiding malware in model files or using special trigger words to bypass security. Companies are testing different ways to defend against these attacks, such as checking tool descriptions more carefully and requiring human approval for risky actions. However, experts warn that attackers can still succeed with a small number of targeted attempts, so ongoing monitoring is important.

US Government Tests AI Models from Microsoft, Google, xAI for Security

US Government Tests AI Models from Microsoft, Google, xAI for Security

The U.S. government may test AI models from Microsoft, Google, xAI, OpenAI, and Anthropic for national security risks before they are released to the public. The Center for AI Standards and Innovation (CAISI) says these tests look for cybersecurity problems, misuse risks, and unpredictable behaviors. Some models are tested with fewer safety limits so experts can see how bad the worst-case situations might be. This program does not require government approval before release, but it might become required in the future. The main risks being checked include cyberattacks and unexpected failures, and lessons learned may help set new safety standards.

Salesforce Pivots AI Pricing to Seat-Based Model After Outcome Test

Salesforce Pivots AI Pricing to Seat-Based Model After Outcome Test

Salesforce talked about charging customers based on the outcomes their AI creates, but found this hard to measure and scale. By late 2025, they said charging by the number of users (seat-based pricing) was becoming more common for their AI products. Industry trends suggest that mixed models - using both subscriptions and outcome-based fees - are growing, but measuring outcomes and making contracts can be difficult. Salesforce's CEO said the company is still figuring out the best way to price AI and may keep offering different options to customers.

CEPI Adopts AI to Speed Pandemic Vaccine Development, Hits Beta in 2026

CEPI Adopts AI to Speed Pandemic Vaccine Development, Hits Beta in 2026

CEPI is using artificial intelligence to make pandemic prediction and vaccine development much faster with its Pandemic Preparedness Engine (PPE). By 2026, the PPE may help researchers find vaccine targets in minutes or days instead of months, but there are still technical and data-sharing challenges. Funding comes from governments, organizations, and grants, and CEPI is working with partners in different countries. Some experts say making a vaccine in 100 days might still be hard for some diseases. New AI tools and safety rules are being tested, and more progress may be seen by 2027 as these systems are built and checked.

Anthropic nears $300M acquisition of dev tools startup Stainless

Anthropic nears $300M acquisition of dev tools startup Stainless

Anthropic is in advanced talks to buy the developer tools startup Stainless for at least $300 million, according to several reports. The deal may be more strategic than just about size, since Stainless helps developers and AI agents connect to top AI models faster and is used by companies like OpenAI, Google, and Anthropic itself. Acquiring Stainless might give Anthropic more control over important developer tools that its rivals also use and could offer better insight into how its models are accessed. The final terms, including payment details, are still being worked out, and no official announcement has been made yet. The $300 million price appears to be a minimum, not a final number, and Stainless could still attract other buyers if the deal changes.

SAP Unveils Autonomous Enterprise Suite with AI Agents at Sapphire 2026

SAP Unveils Autonomous Enterprise Suite with AI Agents at Sapphire 2026

SAP announced its Autonomous Enterprise suite at the Sapphire 2026 conference. This suite combines a Business AI Platform with agents embedded in core applications, which may help automate processes in areas like finance, supply chain, human resources, and customer experience. The new suite includes over 50 assistants and more than 200 specialized agents designed to run end-to-end workflows, though its real-world effectiveness still needs to be validated. SAP emphasizes built-in governance and audit trails to manage risk, but pricing details have not been disclosed. Experts suggest this may help existing SAP clients upgrade automation without major custom changes, though some uncertainty remains until it is tested at scale.

DX upgrades raise AI productivity by 10% for dev teams

DX upgrades raise AI productivity by 10% for dev teams

Studies suggest that improving developer experience (DX), like better tools and clear processes, may raise AI productivity in developer teams by about 10%. Teams with good DX see faster workflows and more reliable AI agents. Surveys show most engineers use AI helpers, but actual time savings are limited unless bottlenecks like slow tests are fixed. Clear command-line tools and well-written documentation also appear to help both humans and AI work more smoothly. Continued improvements to DX, not just new AI models, may be key to ongoing productivity gains.

CME Group and Silicon Data Launch GPU Futures Market

CME Group and Silicon Data Launch GPU Futures Market

CME Group and Silicon Data announced a plan to create a futures market for GPU computing power, letting traders lock in prices for standardized units based on daily rental rates. The contracts may help businesses manage costs as GPU prices have swung by over 30 percent, but the project still needs approval from regulators and has some technical details to resolve. Launch timing and trading volumes remain uncertain, and experts suggest more contract types may be needed due to differences in hardware and regions. Early users might include cloud providers, asset managers, and data-center operators who want stable prices. Standardizing GPU pricing appears difficult because of price differences across locations and hardware, and no test trades have taken place yet.

Google Blocks First AI-Generated Zero-Day Exploit in May 2026

Google Blocks First AI-Generated Zero-Day Exploit in May 2026

In May 2026, Google stopped the first known case where criminals used AI to create and try to use a new security flaw, or zero-day exploit. The attackers made a Python script that targeted a problem in two-factor authentication, but Google caught it before it caused big problems. Experts say this case suggests that AI tools may help criminals find and use security flaws much faster than before. There may be more AI-made attacks like this happening already, and defenders are now looking at using AI to help find and fix these issues quickly. Experts also warn that human oversight is still needed to avoid mistakes from trusting AI too much.

S&P 500 AI mentions surge, but productivity gains remain elusive

S&P 500 AI mentions surge, but productivity gains remain elusive

Mentions of AI by S&P 500 companies have reached record highs, but clear productivity gains have not yet appeared. Executives often express hope about doing more work with the same number of staff, but hard data is still limited, so predictions rely mostly on outside studies. Some reports suggest AI might affect millions of jobs, but opinions about whether jobs will be gained or lost are not settled. There is strong demand for AI skills and retraining, while routine jobs may shrink. Whether recent talk about AI will lead to real cost savings is still uncertain and will be clearer after more data comes in.

Anthropic unveils Managed Agents upgrades, partners with SpaceX Colossus

Anthropic unveils Managed Agents upgrades, partners with SpaceX Colossus

Anthropic announced new features for its Managed Agents and a partnership with SpaceX's Colossus supercomputer. These upgrades may help agents run all the time without needing people to watch over them. The Managed Agents now include tools for working in teams, learning from past sessions, and measuring progress. Anthropic also gained access to over 220,000 GPUs from SpaceX to handle more traffic. This suggests Anthropic wants to make it easier for businesses to use and manage AI agents at a large scale.

Pasadena Chamber expands webinar lineup, adds AI training

Pasadena Chamber expands webinar lineup, adds AI training

The Pasadena Chamber is adding more webinars, including sessions about a Japan trip, tariff refunds, and using AI in business. The AI webinar may help companies understand both the benefits and limits of automation, with topics such as chatbots and bookkeeping. There is also a special focus on small businesses, as many appear to be interested in using AI tools, though some workers reportedly worry about its impact on company reputation. The Chamber's schedule includes over 30 events in May, both online and in-person. Businesses might be able to get tariff refunds if they meet certain conditions set after a court ruling, but there could be waiting times and taxes on the refunds.

US government expands early AI model testing with Microsoft, Google, xAI

US government expands early AI model testing with Microsoft, Google, xAI

The US government has made agreements with Microsoft, Google, and xAI to test new AI models for national security risks before they are released to the public. This program appears to focus on finding possible problems like hacking, dangerous chemicals, or loss of human control, while companies can still make changes. The testing is voluntary, and the government is not forcing companies to join or change their products. These steps may suggest a shift toward checking AI tools before they cause real-world issues. Experts think this ongoing testing and feedback might help shape future rules, but right now it is described as a partnership, not a requirement.