OpenAI is expanding its response to child sexual abuse material (CSAM) while AI-related reports surge. Childlight says reports of AI-generated abuse material climbed from about 4,700 in 2023 to more than 400,000 in 2025, and estimates more than 300 million children worldwide experience at least one form of technology-related sexual exploitation or abuse each year. It urges governments to treat child sexual exploitation as a preventable public health emergency.
In the United States, Western North Carolina prosecutors have used older obscenity laws to pursue AI-generated CSAM cases because, U.S. Attorney Russ Ferguson said, the law has not kept pace with the technology. So far, people prosecuted by his office in such cases also possessed real CSAM. Ferguson supports a specific ban on AI-generated images of children being sexually abused, while acknowledging it could face First Amendment challenges. OpenAI, meanwhile, offers vetted organizations an image-scoring tool at no cost through its API. Early testing suggests it can reduce images sent to human reviewers by 48% while detecting about the same amount of abusive material. OpenAI is also adding ChatGPT warnings and 24-hour suspensions for attempts to find such material, with longer restrictions or permanent account closure possible for repeat violations.
Hamilton Lane is stepping further into venture investing amid the AI boom. The private-markets firm held a $370 million first close for its latest venture fund in 2026, after closing a $615 million fund in early 2025. Executives described the strategy as finding “very attractive value”; the latest fund’s target was not disclosed.
AI-linked markets are also showing volatility. Leopold Aschenbrenner’s hedge fund, Situational Awareness, held shares in Australian neocloud company SharonAI and options in neocloud firms at the end of the third quarter. By the end of October, SharonAI had fallen 58% from September 30, while IREN was down 41% and Nebius 35%. The filing does not show whether the fund sold during the decline; it had nearly collapsed over the summer and sold its public stock positions to Ken Griffin’s Citadel. Separately, DigitalOcean shares fell 6.4% after it reported a wider first-quarter net loss of $43.8 million, despite 24% year-over-year revenue growth. SoundHound AI, Oracle, MongoDB and Fastly also fell amid concerns about rates and the durability of the AI rally; Oracle, MongoDB and Fastly reported weaker new software licenses and slower-than-expected overall revenue growth despite cloud growth.
Companies are also positioning for AI-related demand and adoption. South Korea’s RF Materials reported nine-month revenue of KRW 69,837 million and operating income of KRW 13,357 million, with communications packages and military components driving sales. It plans capacity growth, supported by a rights offering and investment in a new site, as demand from AI data centers rises. In Japan, StrikeReady and Sanwa Comtec are offering StrikeReady’s security platform, which uses Composite AI across alerts, incidents, vulnerabilities, assets, identities and threat intelligence. The companies say it works with existing security tools, has no usage limits tied to spending, and cut incident investigation time by 80% in one enterprise deployment.
OpenAI is bringing computer-use agents into healthcare software. The agents can read screens, click, type and move between applications; early users include Abridge, Ambience, Cigna Healthcare and CVS Health, for tasks such as referrals and document handling. OpenAI says GPT-6 Astra received safe ratings on 99% of 325 reviewed clinician inbox tasks, and doctors rated it very or extremely helpful in 93% of 646 assessments. CVS Health said one workflow cleared a months-long document-sending backlog over a weekend and released an Agent Kit built with OpenAI’s Agents SDK. OpenAI says customers control app access and which tasks need human review, but did not disclose agent model names, error rates or healthcare workspace pricing.
Outside corporate AI deployments, the Guernsey Institute is asking employers what skills workers will need, in its first such survey in 12 years. Institute leaders expect AI to feature prominently and say faster-changing skills make workforce preparation and opportunities for further learning important. In China, October industrial output grew 4.8% year over year, but retail sales rose only 0.4% and property investment fell 21.6% over the first 10 months. Strong exports and advanced manufacturing are supporting growth despite weak domestic demand, potentially allowing policymakers to rely on targeted rather than broad spending support. InvestingPro is also promoting discounted research subscriptions, including monthly AI stock picks, 88 AI-powered strategies and analysis of more than 60,000 stocks; it cautions that examples of returns are not representative of all picks and past performance does not guarantee future results.
Key Takeaways
- Childlight reports that AI-related CSAM reports rose from about 4,700 in 2023 to more than 400,000 in 2025, and estimates over 300 million children experience technology-related sexual exploitation or abuse annually.
- Western North Carolina prosecutors have used older obscenity laws in AI-generated CSAM cases; U.S. Attorney Russ Ferguson backs a specific ban but notes possible First Amendment challenges.
- OpenAI’s image-scoring tool may cut images sent for human review by 48% while detecting about the same amount of abusive material; vetted organizations can access it free through the API.
- OpenAI is adding ChatGPT warnings and 24-hour suspensions for attempts to find CSAM, with longer restrictions or permanent closure possible for repeat violations.
- Hamilton Lane held a $370 million first close for its latest venture fund in 2026, following a $615 million fund close in early 2025.
- Situational Awareness held SharonAI shares and neocloud options at the end of Q3; by late October SharonAI was down 58%, IREN 41% and Nebius 35% from September 30.
- DigitalOcean fell after reporting a wider first-quarter net loss of $43.8 million, although revenue grew 24% year over year; several other software and AI-related stocks also declined.
- RF Materials reported nine-month revenue of KRW 69,837 million and operating income of KRW 13,357 million as it plans capacity expansion amid AI data-center demand.
- StrikeReady and Sanwa Comtec are bringing StrikeReady’s AI security platform to Japan; one enterprise deployment cut investigation times by 80%, according to the company.
- OpenAI healthcare agents are being tested by Abridge, Ambience, Cigna Healthcare and CVS Health; CVS says an agent workflow cleared a months-long document backlog over a weekend.
AI child abuse reports surge more than 8,400% in two years
Childlight, a research institute based in Edinburgh, found that reports of AI-related child sexual abuse material rose from about 4,700 in 2023 to more than 400,000 in 2025. The material can include fake images made with “nudify apps,” which may be used to blackmail children. Childlight estimates that more than 300 million children worldwide experience at least one form of technology-related sexual exploitation or abuse each year. Its research also found that almost a third of girls and more than one in 10 boys in North America experienced in-person sexual abuse during childhood. Childlight urged governments to treat child sexual exploitation as a preventable public health emergency.
Prosecutors turn to older laws to pursue AI child abuse cases
U.S. Attorney Russ Ferguson said on Nov. 13 that prosecutors in Western North Carolina have encountered cases involving AI-generated child sexual abuse material. He said the law has not kept pace with the technology, so prosecutors have used older obscenity laws. In all such cases prosecuted by his office so far, the accused also possessed real child sexual abuse material. Ferguson supports a law that specifically bans AI-generated images of children being sexually abused, though such a law could face First Amendment challenges. David Tatum, a Charlotte child psychiatrist who used AI to alter images of minors, was sentenced to 40 years in prison in 2023.
OpenAI expands tools to detect and discourage child abuse material
OpenAI is offering a tool that scores potentially abusive images so human reviewers can prioritize cases. The company says early testing suggests it could send 48% fewer pictures for review while detecting about the same amount of abusive material. Fal reported that reviewers confirmed 72% of high-confidence flags as previously unidentified child sexual abuse material, while SafetyKit said the tool detected 90% of cases involving drawings or animations that sexualize children in its evaluation. OpenAI is also adding ChatGPT warnings and 24-hour account suspensions for attempts to find such material, with longer restrictions or permanent closure possible for repeated violations. The company says it developed the image classifier with help from the National Center for Missing & Exploited Children and offers it at no cost to vetted organizations through its API.
Hamilton Lane tracks AI boom as it seeks value in venture investing
Private-markets firm Hamilton Lane is tracking the AI boom through its venture strategy, which executives described as finding “very attractive value.” The firm held a first close for its latest fund at $370 million earlier in 2026, executives said during a November earnings call. Hamilton Lane declined to comment on the fundraising, and the fund’s target could not be learned.
AI boom drives Hamilton Lane further into venture investing
Hamilton Lane is raising its second venture fund after launching a dedicated venture strategy in 2023. The private-markets firm held a $370 million first close for its latest fund, following the closing of a $615 million fund in early 2025. The article identifies the firm’s move as part of its deeper involvement in venture investing amid the AI boom.
Aschenbrenner’s fund returned to AI stocks before October losses
Leopold Aschenbrenner’s hedge fund, Situational Awareness, bought AI-related stocks and options as it returned to markets near the end of the third quarter. A regulatory filing showed a stake in Australian neocloud company SharonAI, along with options in neocloud companies. By the end of October, SharonAI had fallen 58%, IREN 41%, and Nebius 35% from September 30. The filing only shows the fund’s holdings at quarter end, so it is not known whether Aschenbrenner sold during the October selloff. The fund had nearly collapsed over the summer and sold its public stock positions to Ken Griffin’s Citadel.
RF Materials reports higher earnings as it expands capacity for AI demand
RF Materials reported 9M revenue of KRW 69,837 million and operating income of KRW 13,357 million. Communications packages and military equipment components drove sales. The company is supporting planned capacity growth with a rights offering and investment in a new site, as demand from AI data centers rises.
Guernsey education leaders ask employers what skills students will need
The Guernsey Institute has launched a survey asking employers what skills workers will need in the future. Executive principal Dr Louise Misselke said it was the first survey of its kind in 12 years and expected artificial intelligence to feature strongly. Simon Le Tocq, chief executive of the institute’s business school, said skills are becoming outdated more quickly and Guernsey must compete through people’s abilities. Esther Reid, dean of TGI University Centre, said the results could help people return to the island for career changes or further learning.
China’s factories grow as weak spending raises questions about growth
China’s industrial output grew 4.8% year on year in October, while retail sales rose just 0.4%. Property investment fell 21.6% over the first 10 months of the year, according to official data. The article says strong exports and advanced manufacturing are helping support growth despite weak domestic demand. This could allow policymakers to rely on targeted support instead of a broad increase in spending.
StrikeReady and Sanwa Comtec bring unlimited AI security tools to Japan
StrikeReady has partnered with Japan’s Sanwa Comtec to offer its security platform to organizations in Japan. The platform uses Composite AI to cover alerts, incidents, vulnerabilities, assets, identities, and threat intelligence, working with companies’ existing security tools. StrikeReady says its AI coverage has no usage limits tied to spending and that customers in one enterprise deployment cut incident investigation times by 80%. The companies are exploring integrations with selected Japanese security solutions and are already evaluating use cases with prospective customers. The service is available in Japan through Sanwa Comtec.
Tech stocks fall as investors question AI rally and rate risks
DigitalOcean, SoundHound AI, Oracle, MongoDB, and Fastly shares fell during afternoon trading as investors worried about rising interest rates and the long-term strength of the artificial intelligence rally. DigitalOcean fell 6.4% after reporting a wider first-quarter net loss of $43.8 million, though revenue grew 24% year over year. SoundHound AI fell 5.5% after announcing an automaker partnership without a significant revenue commitment. Oracle, MongoDB, and Fastly dropped after each reported a decline in new software licenses and slower-than-expected overall revenue growth, even as their cloud businesses grew.
InvestingPro offers AI stock picks and research tools at up to 55% off
InvestingPro is offering its research tools at a discount during earnings season, including monthly updated AI stock picks and tools for reviewing company results, valuing shares, and studying charts. The article lists examples of stock picks and their reported returns, including Ardent Health, Yubico AB, and Kingboard Holdings. It says these are live results, not backtests, and warns that the examples do not represent all picks and that past performance does not guarantee future results. Members also get 88 AI-powered strategies across global markets and analysis of more than 60,000 stocks.
OpenAI brings computer use agents to hospital software
OpenAI said its computer-use agents can read screens, click, type, and move between applications to handle work in healthcare software. Early users include Abridge, Ambience, Cigna Healthcare, and CVS Health, which use agents for tasks such as referrals and document handling. OpenAI reported that GPT-6 Astra earned safe ratings on 99% of 325 reviewed clinician inbox tasks, while doctors rated it very or extremely helpful in 93% of 646 assessments. CVS Health said one workflow cleared a months-long document-sending backlog over a weekend, and it released an Agent Kit built with OpenAI’s Agents SDK. OpenAI said users can control which apps agents access and which tasks require human review; it did not provide agent model names, error rates, or healthcare workspace pricing.
Sources
- Reports of AI-related child sexual abuse material rise 8,400% in two years, data shows
- Prosecutors use old laws to fight AI child abuse images
- OpenAI child safety tools target AI-generated abuse
- Hamilton Lane’s Venture Strategy Is Tracking the AI Boom. Here’s Where It’s Finding ‘Very Attractive Value’
- AI Boom Pulls Investment Firm Hamilton Lane Deeper Into Venture
- Aschenbrenner’s Situational Awareness made AI bets again before October rout
- RF Materials: Stronger 9M earnings and communications-package growth accompany capacity investment for AI-driven demand
- Students must adapt to AI, says Guernsey education boss
- China’s AI Export Engine Kept Humming As Spending Stayed Soft
- StrikeReady Expands to Japan with Sanwa Comtec, Bringing Unmetered AI to Security Teams Stretched Thin
- DigitalOcean, SoundHound AI, Oracle, MongoDB, and Fastly shares are falling, what you need to know
- Research earnings season with InvestingPro AI stock picks and tools for up to 55% off
- OpenAI pushes agents into hospital software
Comments
Please log in to post a comment.