President Trump announced a voluntary safety accord with top AI leaders at the White House on Thursday. The agreement relies on self-regulation by companies rather than new government rules. Trump emerged from a closed-door meeting with prominent industry executives standing behind the accord, which gives companies responsibility for their own controls and audits without imposing new sanctions or transparency obligations.
Despite the voluntary nature of the deal, nearly three-quarters of senior Hill staffers said Congress should act to regulate the AI industry this year, according to a Canvass survey. The push highlights growing concern among congressional aides about the need for oversight even as the administration embraces a lighter-touch approach.
OpenAI disrupted a coordinated campaign that tried to extract protected reasoning from its models without permission. The activity began July 1 and included 16,000 requests from over 4,000 users during high-volume spikes. OpenAI attributes a core cluster to individuals linked to Moonshot AI, maker of Kimi, and has banned fraudulent accounts while strengthening protections for hidden reasoning.
Bank of America stock may be 39 percent undervalued as AI spending rises. Management is increasing AI spending on platforms like Erica and tools such as AskGPS and CashPro, and plans to double its AI budget next year to support efficiency and fee income growth.
Denny's CEO Christopher Bode announced plans to upgrade roughly 500 restaurants over the next 12 to 18 months as part of Project Grand Slam. AI-powered tools will help set menu prices based on local demand, and the chain is also rolling out new point-of-sale systems, table-side tablets, and an upgraded mobile app.
IBM's global study finds most CFOs now help lead enterprise AI strategy, and by 2030 over half expect to design financial and ethical guardrails for AI. AI-first CFOs help their organizations grow revenue 23 percent faster and approve AI funding 15 percent quicker than peers. Arch expanded its AI platform with Investment Research, a tool for private markets diligence that early partners say reduced diligence time by up to 50 percent per opportunity.
Researchers introduced DerivAudit, a framework to check whether AI memories are supported by the history available when written. Broader evidence recovers support for about 60 percent of memories that appear unsupported from citations alone. Separately, BudgetPM allocates observations under a shared budget, with BudgetPM-Static cutting observations by 42 to 54 percent while keeping nearly all quality.
Cerebras Systems share price is down 37.3 percent year to date despite a new collaboration with Gimlet Labs to bring Cerebras hardware into Gimlet Cloud. Its price-to-book ratio of 5.1x sits above the semiconductor sector average of about 4.5x. Brookfield Infrastructure Partners targets 10 percent FFO growth driven by its organic backlog and recent investments in renewable energy and transportation.
Key Takeaways
- Trump signed a voluntary AI safety accord with tech leaders at the White House, relying on self-regulation instead of new government rules
- Nearly 75% of senior Hill staffers want Congress to regulate AI this year, contrasting with the voluntary approach
- OpenAI disrupted a distillation campaign with 16,000 requests from over 4,000 users, linked to Moonshot AI
- Bank of America plans to double its AI budget next year; stock may be 39% undervalued amid AI spending
- Denny's CEO Christopher Bode plans to upgrade 500 restaurants using AI tools for menu pricing as part of Project Grand Slam
- IBM study shows AI-first CFOs grow revenue 23% faster and approve AI funding 15% quicker than peers
- Arch's new Investment Research tool reduced diligence time by up to 50% for early partners in private markets
- DerivAudit framework finds broader evidence supports about 60% of AI memories that appear unsupported from citations alone
- BudgetPM-Static cuts observations by 42 to 54% while maintaining nearly all quality in AI agent systems
- Cerebras stock down 37.3% year to date despite new Gimlet Labs collaboration; Brookfield targets 10% FFO growth
Trump and AI leaders sign voluntary safety accord
President Trump announced a voluntary agreement with top AI leaders at the White House on Thursday. The accord aims to promote safe and responsible development of AI. It relies on self-regulation by companies rather than new government rules.
Trump unveils voluntary AI safety accord with tech leaders
President Donald Trump and leading technology executives announced a new voluntary industry agreement on AI safety. The agreement emphasizes self-regulation by companies instead of new government rules. Trump emerged from a closed-door meeting with prominent industry leaders standing behind the White House accord.
Trump and Big Tech make AI security voluntary
The White House agreement gives companies responsibility for controls and audits. It does not impose new sanctions or transparency obligations. President Trump has rebranded AI as superintelligence, but the legal substance remains unchanged.
Bank of America stock may be undervalued amid AI spending
Bank of America stock may be 39 percent undervalued as AI spending rises. Management is increasing AI spending on platforms like Erica and tools such as AskGPS and CashPro. The bank plans to double its AI budget next year to support efficiency and fee income.
Cerebras Systems stock looks pricey despite AI news
Cerebras Systems share price is down 37.3 percent year to date. The company has a new collaboration with Gimlet Labs to bring Cerebras hardware into Gimlet Cloud. Its price-to-book ratio of 5.1x sits above the semiconductor sector average of about 4.5x.
IBM study shows CFOs take bigger AI leadership roles
A global IBM study finds most CFOs now help lead enterprise AI strategy. By 2030, over half expect to design financial and ethical guardrails for AI. AI-first CFOs help their organizations grow revenue 23% faster and approve AI funding 15% quicker than peers. The study calls for connecting strategy, governance, and capital allocation to create value.
Auditing AI memory for unsupported facts
Long-running AI agents store past conversations as memory, but that memory may not match the original evidence. Researchers introduce DerivAudit, a framework to check whether memories are supported by the history available when written. Broader evidence recovers support for about 60% of memories that appear unsupported from citations alone. However, expanding evidence does not make admission fully reliable.
Budgeting agent observations for stored intentions
AI agents can store intentions for future conditions, but checking those intentions costs resources. Researchers introduce BudgetPM, a system that allocates observations under a shared budget. BudgetPM-Static uses a Logistic scorer and keeps nearly all quality while cutting observations by 42 to 54 percent. BudgetPM-Sequential performs better under scarcity by deciding when to spend or reserve capacity.
Arch launches AI tool for pre-investment diligence
Arch expanded its AI platform with Investment Research, a tool for private markets diligence. It helps investment teams assess opportunities by pulling key terms from offering documents and flagging items for review. Early partners reduced diligence time by up to 50 percent per opportunity and saved thousands of hours. The tool can be bought alone or combined with Arch's data management and reporting tools.
OpenAI disrupts coordinated AI model distillation campaign
OpenAI disrupted a campaign that tried to extract protected reasoning from its models without permission. The activity began July 1 and included 16,000 requests from over 4,000 users during high-volume spikes. OpenAI attributes a core cluster to individuals linked to Moonshot AI, maker of Kimi. The company banned fraudulent accounts, strengthened protections for hidden reasoning, and shared findings with industry partners.
Hill staffers urge Congress to regulate AI industry this year
Nearly three-quarters of senior Hill staffers said Congress should act to regulate the artificial intelligence industry. The findings come from a latest Canvass survey. The push highlights growing concern among congressional aides about the need for oversight.
Denny's plans upgrades for 500 restaurants using AI technology
Denny's CEO Christopher Bode announced plans to upgrade roughly 500 restaurants over the next 12 to 18 months. The effort is part of Project Grand Slam, a five-part turnaround strategy. AI-powered tools will help set menu prices based on local demand, and the chain is also rolling out new point-of-sale systems, table-side tablets and an upgraded mobile app.
Brookfield Infrastructure targets 10% growth as AI investments grow
Brookfield Infrastructure Partners said it expects its growth rate to accelerate. The company targets 10% FFO growth driven by its organic backlog and recent investments. Management stressed a commitment to financial discipline while increasing investments in areas including renewable energy and transportation.
Sources
- Trump, top AI leaders agree to voluntary "accord" on AI standards
- 'Tremendous self-policing': President Trump unveils voluntary AI safety accord with tech leaders
- Trump poses for a photo with Big Tech; AI security remains voluntary
- Bank Of America (BAC) Stock May Be 39% Undervalued As AI Spending Rises
- Cerebras Systems (CBRS) Stock Looks Pricey Despite Fresh AI Inference News
- IBM Study: As AI Scales Enterprise-Wide, CFOs Play an Expanded Role in Transformation
- Memory Is a Derivation: The Distributed-Evidence Paradox in Long-Term Agents
- When Should Agents Check External State? Budgeting Observations for Stored Intentions
- Arch extends AI platform into pre-investment diligence
- Disrupting a coordinated model-distillation campaign
- Hill aides: AI industry needs regulation this year
- Denny's targets 500 restaurant upgrades as diner chain invests in AI, new technology: 'Project Grand Slam'
- Brookfield Infrastructure Targets 10% FFO Growth as AI Investments Accelerate
Comments
Please log in to post a comment.