Week in Review (11/2/24)

It’s hard to keep up with the constant stream of AI software updates.  Here are some highlights from last week:

• OpenAI Desktop Voice: ChatGPT’s Voice mode with nine voices now on Mac/Windows for Plus users, supporting natural voice interactions.
• Anthropic Desktop App: Claude AI now on desktop for Mac/Windows with mobile voice dictation, improving access.
• OpenAI ChatGPT Search: New real-time search delivers live answers with sources on desktop and mobile.
• Apple AI: New “Apple Intelligence” tools on select iPhones enhance messaging, notifications, and image editing.

This was also a big week with the 1-year anniversary of the White House’s AI Executive Order.  Over the last year, there have been over 100 actions completed to advance AI safety, privacy, and U.S. AI leadership.  Great progress!

Here's the rest of this week's AI news.

Federal

• Dept of Education releases AI toolkit to guide responsible use of AI in education The Department of Education’s new AI toolkit offers step-by-step guidance on safely integrating AI into education, focusing on privacy, strategy, and equitable practices for diverse learning environments.
• Air Force names Susan Davenport as new chief data and AI officer Susan Davenport, a government tech expert, has been named Chief Data and AI Officer for the Air Force, tasked with driving AI-readiness by 2025 and implementing enterprise-level AI initiatives.
• Nearly 60 militaries now endorse U.S.-launched pledge on ‘responsible AI’ Nearly 60 nations endorse the U.S.-led pledge for responsible AI in military use, aiming to establish ethical standards. The Pentagon’s forum gathers experts for cross-sector discussions on AI policy.
• FTC on track to publish its first AI use case inventory by the end of this year The FTC aims to release its initial AI use case inventory by year-end, with a focus on responsible AI governance within the agency’s broader data governance framework.
• DOD tests AI software, advances to improve physical security posture The Department of Defense tested Scylla, an AI security software, demonstrating its high detection accuracy for physical security threats, enhancing surveillance and reducing false alarms at key military sites.
• AI ‘will enhance’ nuclear command and control, Says STRATCOM boss STRATCOM aims to modernize nuclear command and control with AI enhancements for rapid, secure data processing, while ensuring human oversight, says Gen. Anthony Cotton.
• HHS’ AI plans come into view HHS is forming an AI regulatory plan across its agencies, focusing on compliance and safe AI in healthcare, while considering industry partnerships for audits and private compliance labs.
• Biden-Harris Administration marks one year of groundbreaking AI regulation progress The Biden-Harris Administration celebrates major AI milestones, including over 100 actions to enhance AI safety, privacy, and civil rights, foster AI research, and strengthen U.S. global AI leadership.
• DHS completes AI pilots, advances AI corps to bolster national security and infrastructure DHS reports successful completion of AI pilot programs enhancing USCIS training, HSI investigations, and FEMA resilience, alongside hiring 31 AI Corps experts to further strengthen critical infrastructure and cybersecurity defenses.
• Making AI work inside and out at SSA Social Security Administration CAIO Brian Peltier shares insights into AI's role in SSA, emphasizing responsible governance, team growth, and the importance of AI literacy to enhance agency operations.
• DHS touts AI work and pilots a year after executive order DHS advances AI capabilities with successful pilot programs for USCIS, HSI, and FEMA, expands AI Corps, and prioritizes transparency and security in AI deployment through compliance and open-source standards.
• Eyeing ‘AI at scale,’ intel community aims to get data house in order ODNI introduces a "data reference architecture" to streamline data handling across intelligence agencies, supporting scalable AI initiatives and fostering collaboration with Five Eyes partners and industry stakeholders.
• AI’s future depends on who’s at the table — not just who’s in the Oval Office AI policy success hinges on startup inclusion in regulatory discussions, argues Gaurab Bansal, as Big Tech’s dominance risks stifling innovation critical for responsible and diverse AI advancements.
• VA announces creation of new AI testing ground with FDA The VA and FDA announced a virtual lab to test healthcare AI technologies, ensuring alignment with federal standards on safety, effectiveness, and adherence to trustworthy AI principles.
• Americans want AI disclaimers, FEC commissioner says FEC Commissioner Shana Broussard highlighted public support for disclaimers on AI-generated political ads, emphasizing the need for Congress to pass legislation to mandate these transparency measures.
• Commerce marks on-year anniversary of AI EO Marking a year since President Biden’s AI Executive Order, Commerce Secretary Gina Raimondo highlights achievements, including the AI Safety Institute’s launch, interagency testing, and international collaborations for responsible AI.
• How DOD will help agencies comply with the White House’s new rules for AI in national security Acting Deputy CDAO for Policy John Turner shares how DOD will lead interagency AI policy compliance, utilizing Responsible AI toolkits and refining national security practices for AI-enabled systems.
• How will GenAI change cybersecurity teams? Nearly half of government cybersecurity professionals expect generative AI to eliminate certain skills and roles, yet 67% call for stronger regulations to safely integrate AI within cybersecurity
• Nearly 60 militaries now endorse U.S.-launched pledge on ‘responsible AI’ Nearly 60 countries now endorse the U.S.-led Political Declaration on Responsible Military AI, with the Pentagon’s CDAO advancing international RAI standards through a closed forum.
• The federal government wants to teach workers about AI prompt engineering The Federal Acquisition Institute’s new AI Prompt Engineering Credential aims to equip federal acquisition staff with practical skills to evaluate and optimize large language models for government use
• A government AI lab is born The VA and FDA have launched a virtual lab to evaluate healthcare AI tools, aiming to standardize safety and effectiveness protocols across the U.S.'s largest health system.
• State to develop new AI marketplace for staff The State Department is developing an AI marketplace to streamline AI adoption across its bureaus, enhancing diplomatic efforts with secure, customizable tools and evolving capabilities for classified data handling.
• Meta is pushing for the government to use its AI Meta’s CEO announced plans to work with the U.S. government to implement the Llama AI model, with details on specific uses and agencies involved remaining undisclosed.

State/Local

• False citations show Alaska education official relied on generative AI Alaska’s education department posted AI-generated policy drafts with fabricated citations, highlighting risks in government AI use and the need for transparency and robust verification protocols.
• NYC council committee considers more transparency around AI NYC's AI progress report shows steady action plan progress, but councilmembers urge enhanced transparency with proposed legislation requiring updated public lists of approved AI tools for city use.
• Miami-Dade County, Fla., works to put AI in action At the South Florida Digital Government Summit, Miami-Dade County officials detailed AI initiatives supporting workforce upskilling, small business AI access, and an AI-powered app to match pets with new owners.
• Values drive the work of AI groups in Washington, D.C. Washington, D.C.’s AI Advisory Group and Taskforce advance AI projects that align with values set by Mayor Bowser, including tools for data accessibility and employment support through “Career Ready DC.”
• SmartCitiesWorld and Microsoft report highlights AI's transformative impact on cities The "AI in Cities Trend Report 2024" explores AI applications in urban areas, such as citizen experience personalization and productivity enhancements, spotlighting Norway's Stavanger Kommune's real-time healthcare dashboards.
• Alabama AI Task Force nearing completion of policy recommendations for state government Alabama’s AI Task Force will submit policy recommendations to the governor, focusing on data management, ethical use, and privacy protocols for AI across state government systems.
• AI takes center stage at Texas conference, exploring transformative uses across government Texas DIR's Discover Conference highlighted AI's rising role in state operations, from cybersecurity to workforce training, with Texas advancing an AI Innovation Lab and legislative efforts on AI ethics.

International

• The emerging age of AI diplomacy U.S. policymakers balance AI collaboration with Gulf states against risks of China’s influence, focusing on cautious technology sharing and investment safeguards to advance American interests in the Middle East.
• Chinese AISI counterparts The US and UK AI Safety Institutes inspire "AISI counterparts" in China, such as CAICT and Shanghai AI Lab, focusing on technical safety, standards, and global AI governance cooperation.
• The need for ethical AI leadership Australia’s public desires clear AI standards and transparency in government use, urging ethical leadership to balance AI’s benefits with accountability and fairness concerns.
• Ukraine rolls out dozens of AI systems to help its drones hit targets Ukraine deploys AI-powered drone systems to counteract signal jamming, enhancing accuracy against high-value Russian targets despite electronic warfare barriers.
• Chinese researchers develop AI model for military use on back of Meta's Llama Researchers from China’s PLA-linked institutions repurposed Meta's Llama AI model for military intelligence tasks, despite Meta’s restrictions against military use, sparking scrutiny from the Pentagon.
• Iran ranked 94th in Government AI readiness Oxford Insights ranks Iran 94th for AI preparedness in public service, citing strengths in data infrastructure but noting gaps in governance. Iran leads Islamic countries in AI publications, with major public-private investments planned for AI infrastructure.
• Vietnamese AI firms have many chances of development Vietnam’s AI sector is rapidly expanding, with government incentives, a strong tech ecosystem, and surging domestic demand positioning Vietnamese AI firms to lead in Southeast Asia’s AI market.
• Malaysia plans to launch national AI office Malaysia will open the National AI Office on December 12 to lead AI policy development, foster innovation, and support digital economy growth, according to Digital Minister Gobind Singh Deo.
• AI Safety and the US-China arms race This analysis argues that AI safety measures won't hinder U.S. competitiveness with China, suggesting that AI safety builds public trust and adoption, enhances economic growth, and safeguards national security interests.
• The AI safety institute international network: next steps and recommendations AI safety experts from nine countries and the EU will convene in San Francisco to set collaborative goals for an international network of AI safety institutes, aiming to align safety practices and bolster public trust in AI systems.
• ASIC warns governance gap could emerge in first report on AI adoption by licensees ASIC advises financial licensees to enhance governance to match increasing AI use, highlighting risks like consumer bias, data security, and ethical challenges without proactive oversight.

Other

• Microsoft Copilot AI use extends deep into corporate America, but companies aren’t 100% sold Microsoft 365 Copilot sees widespread corporate adoption, but firms remain cautious on full deployment, citing cost and infrastructure challenges, according to a CNBC survey of tech executives.
• Can startups thrive in an age of AI? AI’s data and resource demands push startups to shift focus from disruption to transformation, prioritizing partnerships with established firms and exploring end-to-end services beyond intermediary software roles.
• Apple's AI features roll out on iPhones - but not for all Apple’s AI suite, “Apple Intelligence,” launches selectively on updated iPhone 15 Pro and iPhone 16 models, offering tools for messaging, notifications, and image editing, with more features expected soon.
• Growing up: navigating generative AI’s early years – AI Adoption Report A new Wharton and GBK Collective report reveals rising Gen AI adoption, with weekly business use nearly doubling since 2023. Companies now emphasize ROI and practical integration for long-term success.
• Pushing the frontiers of audio generation Google's latest speech generation technologies deliver more natural, multi-speaker dialogues with improved acoustic quality, powering features like NotebookLM Audio Overviews, Illuminate, and YouTube's auto-dubbing tools.
• DeepLearning.AI Batch In this Halloween-themed issue of The Batch, Andrew Ng examines the psychology behind AI "fear mongering," noting that speculative warnings often stem from incentives to gain visibility, attract investment, or influence regulation.
• Accenture forms new Nvidia business group focused on agentic AI adoption Accenture's new Nvidia business group leverages AI Refinery to accelerate autonomous, task-focused AI adoption, supported by global engineering hubs and a 30,000-strong Nvidia-trained workforce.
• Google CEO says more than 25 percent of company’s new code written by AI Google CEO Sundar Pichai announced that AI now writes over a quarter of Google’s new code, boosting developer productivity and reflecting the company’s expanded AI investment strategy.
• AI emerging as a not-so-secret business development weapon for government contractors The 2024 GAUGE report highlights AI’s transformative role in GovCon, enhancing bid decisions, proposal quality, and pursuit prioritization for contractors facing fierce competition in revenue growth.
• Leaders in policy and tech call for balanced AI regulation Stanford’s Cyber Policy Center panel highlights calls for balanced AI governance to ensure safety and innovation amid rapid tech growth and challenges like job impacts and digital misinformation.
• Who will be the workers most affected by AI? This study highlights how socio-demographic factors influence workers' access to AI benefits, urging policymakers to tailor supports to foster inclusive productivity and economic growth.
• Anthropic’s Claude AI chatbot now has a desktop app Anthropic's Claude chatbot now offers a desktop app for Mac and Windows, alongside voice dictation for mobile users, enhancing accessibility and convenience.
• OpenAI launches ChatGPT Advanced Voice mode ChatGPT’s Advanced Voice mode, featuring nine voices and responsive interactions, is now accessible on Mac and Windows desktop apps for Plus subscribers, allowing natural, real-time voice conversations.
• Meta makes its MobileLLM open for researchers, posting full weights Meta releases MobileLLM, an optimized language model for mobile devices, under a non-commercial license, allowing researchers to advance on-device AI without cloud dependencies.
• Amazon CEO pledges AI investments will pay off as capital expenditures surge 81% Amazon CEO Andy Jassy reassures shareholders on major AI investments, citing AWS’s growth as a model for returns, with capital expenditures rising 81% to meet AI infrastructure demand.
• Is now the right time to invest in implementing agentic AI? Experts advise caution on investing in agentic AI, recommending a phased approach that complements existing RPA, given the complexity of implementation and early-stage maturity of the technology.
• A toolkit to help you manage uncertainty around AI To manage AI uncertainty, leaders should blend four strategies—thinking, seeing, doing, and shaping—to enhance decision-making and transform uncertainty into a competitive opportunity.
• An executive’s crash course in AI agents AI agents, poised to automate complex workflows, require strategic integration to maximize ROI. Leaders must prioritize adaptive learning and governance to stay competitive amid rapid AI advancements.
• Even Mark Zuckerberg seems surprised by Meta’s pace of spending on AI Meta raised 2024 capital expenditure guidance to $38-40 billion as rapid AI infrastructure expansion exceeded expectations. Despite strong earnings, shares dipped amid concerns over high costs and modest user growth.
• What the past can teach us about our AI fears AI misinformation concerns echo past fears over manipulated media; authors argue that over-regulation could suppress free speech and suggest societal adaptation as a better response.
• Visa has deployed hundreds of AI use cases. It’s not stopping. Visa has deployed more than 500 generative AI tools to enhance productivity and fraud prevention, aiming for quick innovation despite job cuts and challenges in measuring AI's ROI, says Tech President Rajat Taneja.
• Introducing ChatGPT search ChatGPT’s enhanced web search now offers real-time answers with sources, integrating live updates on weather, stocks, sports, and more. Available on desktop and mobile, it aims to boost accessibility to accurate, high-quality content.