1 big thing: AI security startup frenzy

2. Google's Bard launches fact-check features

3. Training data

5. + This

Hi, it's Ryan here, reporting from the permanent traffic jam that is New York City during the UN General Assembly. Today's AI+ is 1,179 words, a 4-minute read.

Thanks to Scott Rosenberg and Meg Morrone for editing and Bryan McBournie for copy editing this newsletter.

A floating, translucent blue 3D render of the human brain.

deepBlue

Taking you inside the AI revolution, and delivering scoops and insights on the technologies reshaping our lives.

Axios AI+

Standard

1 big thing: Microsoft, Google rebuild their worlds around AI

Illustration of an open house window with panes resembling the Microsoft Windows logo. Outside the open window is a bright blue sky filled with glowing binary code.

2. The great AI "pause" that wasn't

4. + This

Ina here. Today's AI+ is 1,284 words, a 5-minute read.

Situational awareness: U.K. regulators have signaled they will not block Microsoft's acquisition of Activision Blizzard ahead of an October deadline to close the deal.

Thanks to Scott Rosenberg and Megan Morrone for editing and Bryan McBournie for copy editing this newsletter.

1 big thing: The right newsroom jobs for AI, and the wrong ones

2. Exclusive: Hurd has GOP primary's first AI plan

LimeWire is back — but this time, it's all about AI

3. LimeWire, now an AI startup, buys BlueWillow

4. Training data

Ina here, happy to be back in your inboxes after spending a week down with a mild case of COVID-19. Today's AI+ is 1,328 words, a 5-minute read.

Axios Pro is out with a new AI policy whitepaper — the legislation, lawmakers and companies to watch right now. Download it here.

1 big thing: AI screens for mental health

Illustration of a hand reaching out darkness towards a hand in a white lab coat reaching out of a computer

2. Big Tech works to speed cancer research with AI

Hi, it's Ryan. Today's newsletter is 1,116 words, a 4-minute read. 

1 big thing: UN deadlocked over regulating AI 

2. Exclusive: How AI policy aligns with voter views

3. OpenAI's updated image generation tool 

4. Training data 

Hi, it's Ryan. Have you read Axios Pro Policy's new white paper on the legislation and lawmakers shaping AI regulation right now? Get on that. Today's AI+ is 1,190 words, a 4-minute read.

Situational awareness: Cisco is buying cybersecurity and data analysis giant Splunk for $28 billion.

National

A new report card from the AI Policy Institute grades the legislative AI proposals that are currently gaining the most traction according to how they line up with U.S. public sentiment. 

Driving the news: A survey of 1,118 voters conducted by YouGov Sept. 2 to 6., is the report card yardstick.

Why it matters: Politicians and wannabe presidential candidates are increasingly touting plans for AI regulation, but those policies don't always align with what voters say they want.

Between the lines: The AI Policy Institute advocates "political solutions to potential catastrophic risks" from AI. Its report card judged proposed AI regulation based on various attributes, including how the proposal adapts to advancements, how it discourages high-risk AI deployment, and how it reduces the proliferation of dangerous AI. 

The AIPI notes that current proposed policies focus on addressing immediate harms as opposed to more long-term existential threats of AI.

Polling shows that voters want future-proof AI regulation that take into account unknown threats. 

The big picture: Americans are consistently telling pollsters they they favor guardrails for AI, especially with regards to election security.

What they're saying: Voters "expect tech companies to be responsible for the products they create" said Daniel Colson, executive director of the AIPI.

Editor's note: This story has been updated with additional detail about AIPI.

Exclusive: How AI policy aligns with what voters want

The AI Policy Institute found that most proposed legislation looks only at current threats.

Artificial intelligence

Automation and AI

A new report card from the AI Policy Institute grades the legislative <a class="gtmContentClick" data-vars-link-text="AI" data-vars-click-url="https://www.axios.com/2023/09/20/open-ai-dall-e-3-image-creator-chatgpt-integration" data-vars-content-id="2755a0c9-f8be-451c-9187-c713e98cd6b6" data-vars-headline="Exclusive: How AI policy aligns with what voters want" data-vars-event-category="story" data-vars-sub-category="story" data-vars-item="in_content_link" href="https://www.axios.com/2023/09/20/open-ai-dall-e-3-image-creator-chatgpt-integration" target="_self">AI</a> proposals that are currently gaining the most traction according to how they line up with U.S. public sentiment. Driving the news: <a class="gtmContentClick" data-vars-link-text="A survey" data-vars-click-url="https://theaipi.org/poll-shows-voters-oppose-open-sourcing-ai-models-support-regulatory-representation-on-boards-and-say-ai-risks-outweigh-benefits-2/" data-vars-content-id="2755a0c9-f8be-451c-9187-c713e98cd6b6" data-vars-headline="Exclusive: How AI policy aligns with what voters want" data-vars-event-category="story" data-vars-sub-category="story" data-vars-item="in_content_link" href="https://theaipi.org/poll-shows-voters-oppose-open-sourcing-ai-models-support-regulatory-representation-on-boards-and-say-ai-risks-outweigh-benefits-2/" target="_blank">A survey</a> of 1,118 voters conducted by YouGov Sept. 2 to 6., is the <a class="gtmContentClick" data-vars-link-text="report card" data-vars-click-url="https://theaipi.org/ai-report-card-09-20/" data-vars-content-id="2755a0c9-f8be-451c-9187-c713e98cd6b6" data-vars-headline="Exclusive: How AI policy aligns with what voters want" data-vars-event-category="story" data-vars-sub-category="story" data-vars-item="in_content_link" href="https://theaipi.org/ai-report-card-09-20/" target="_blank">report card</a> yardstick.

From the Valley to D.C., Big Tech players like Facebook, Google and Amazon are under more scrutiny than ever as new technology develops and privacy and antitrust concerns grow in lockstep with companies’ ambitions.

Technology

If you think the U.S. Congress is moving slowly on AI regulation, you'll be waiting much longer for a global AI regulator or treaty.

The big picture: That's the message out of the UN General Assembly in New York this week, as political leaders, tech companies and civil society gather to debate global challenges.

While leaders including President Biden and UN Secretary-General António Guterres mentioned AI in their set-piece speeches, political gridlock at the UN since Russia invaded Ukraine virtually eliminates the prospect for action. 

Why it matters: A plurality of AI experts surveyed by Axios support global guardrails for AI. 

But in the absence of UN leadership, no organization has asserted authority over the AI safety debates led by groups ranging from the G-7 to the World Economic Forum and the Organization for Economic Cooperation and Development.

Driving the news: The leaders of four of the five permanent members of the UN Security Council skipped this week's debate — only Biden showed up.

A downbeat Guterres this week called for "some global entity" with AI monitoring and regulatory capacity and warned that "governments alone will not be able to tame" AI. But in a CNN interview, he admitted the UN "has no power at all" to bring superpowers together and warned that the world is headed towards a "great fracture." 

Context: Aside from determining what guardrails are needed for powerful AI models, the thorniest question in AI governance is how to involve China.

Most of the international forums that are working on AI governance exclude China by default — including the G-7, OECD, and Council of Europe. 

But the U.K. has invited China to a global AI safety summit it will host in November.

The intrigue: By playing a supporting role in the upcoming U.K. summit rather than seizing control of the global debate, the White House has created space to include China in governance discussions.

In the absence of the UN or U.S. leading the global debate, there's also more room for others traditionally sidelined in global discussions, such as academics, who proposed a range of governance models in a July paper.

The current circular debate leaves executives from tech companies tasked with participating in parallel political dialogues.

Flashback: In the past, the UN has only mobilized to regulate emerging risk — including via UN-affiliated nuclear and civil aviation watchdogs — when spurred to action by catastrophic events.

Yes, but: Effective AI guardrails don't depend on a new global AI agency or treaty. 

The EU is close to finalizing a comprehensive AI law, and advocates of guardrails share common ground around the world. 

What they're saying: Omar Sultan Al Olama, the UAE's AI minister, told Axios that he welcomes any global AI governance effort but "many efforts are non-starters" because they will lack buy-in from developing countries or China. 

"The first thing we need to agree on is principles that cut across" industries and geographies, before leaving national governments to implement those principles, he said.

Aisén Etcheverry Escudero, Chile's technology minister, told those gathered at an AI for Humanity event Wednesday that she wants to frame AI guardrails "not just from a regulation perspective but from a capacity creation perspective," seeing AI as a way to close digital divides globally.

Anthony Aguirre, executive director of the Future of Life Institute, which six months ago led a call for a pause in generative AI development, told Axios he is optimistic that leaders are taking concerns about possible runaway AI seriously. Across multiple forums there's agreement that "people want AI systems that are safe and transparent and unbiased," he told Axios.

What's next: Guterres plans to appoint a multistakeholder advisory board on AI.

a range of governance models in a July paper

catastrophic events to prompt the creation of those bodies

UN deadlocked over regulating AI

That's the message out of the UN General Assembly this week in New York.

If you think the U.S. Congress is moving slowly on AI regulation, you&#x27;ll be waiting much longer for a global AI regulator or treaty.The big picture: That&#x27;s the message out of the UN General Assembly in New York this week, as political leaders, tech companies and civil society gather to debate global challenges.

Republican presidential primary contender Will Hurd is the first candidate to release an AI policy plan — framing coming AI development as a "Cold War with China" and urging licensing of frontier AI models.

Why it matters: Hurd's hawkish, pro-regulation of AI approach is notable given his deep AI credentials.

Hurd, a former Republican member of Congress from Texas, chaired the first AI hearing in the House in 2018. He also claims credit for writing "the first National Strategy for AI" (a congressional resolution) and served on OpenAI's board from 2021 until July.

Hurd writes that "powerful AI models should need to obtain a permit" — and likens them to nuclear power plants.

He wants a Republican AI strategy to "take advantage of technology before it takes advantage of us," and also supports:

"Strict regulations of sensitive AI technology exports."

Ensuring AI is deployed along "every mile of our border." 

Compensating creators "when their creations are utilized in AI-generated content."

Making coding and data analytics standard middle school subjects, noting that "AI threatens 85 million jobs worldwide." 

Between the lines: Though AI promises to shape both 2024 campaign methods and debates, even candidates with tech industry backgrounds — including Vivek Ramaswamy and Gov. Doug Burgum (R-N.D.) — have avoided AI specifics thus far.

Yes, but: Hurd, a moderate who quit his seat in Congress in 2020, has so far failed to qualify for either of the Republican primary debates and has not topped 1% nationally in polls.

Editor's note: This story has been corrected to show that Hurd chaired the first AI hearing in the House of Representatives, not the first AI hearing in Congress. The first AI hearing in Congress was in the Senate in 2016.

first candidate to release an AI policy plan

Exclusive: First AI plan in Republican primary

The former GOP Rep. wants to require licensing for advanced AI models.

Former Rep. Will Hurd (R-Texas).  Photo: Scott Eisen/Getty Images

Republican presidential primary contender Will Hurd is the <a class="gtmContentClick" data-vars-link-text="first candidate to release an AI policy plan" data-vars-click-url="https://www.documentcloud.org/documents/23985202-harnessing-artificial-intelligence" data-vars-content-id="7799e82a-c4f1-4120-922d-d943295ef9c7" data-vars-headline="Exclusive: First AI plan in Republican primary" data-vars-event-category="story" data-vars-sub-category="story" data-vars-item="in_content_link" href="https://www.documentcloud.org/documents/23985202-harnessing-artificial-intelligence" target="_blank">first candidate to release an AI policy plan</a> — framing coming AI development as a &quot;Cold War with China&quot; and urging licensing of frontier AI models.Why it matters: Hurd&#x27;s hawkish, pro-regulation of AI approach is notable given his deep AI credentials.

Users can now ask Google's Bard chatbot to double-check its answers and can connect it to their Google apps and services, in the first update to the product since July.

Why it matters: Google is racing to integrate AI into its main products, hoping to extend their dominance, after being caught out earlier this year when Microsoft integrated ChatGPT into its Bing search engine and added an AI "copilot" to Microsoft 365.

With 1 in 3 Americans describing themselves as "very concerned" about the development of AI — driven partly by the rise of chatbots that cannot explain themselves — citation and transparency improvements could help win over some AI skeptics.

What's happening: New and existing English-language Bard users will be prompted to decide if they want to add Bard Extensions when they use the chatbot.

The opt-in Extensions plug-in connects Bard to Google products such as Gmail, Docs, Drive, Maps, YouTube, and Google Flights. The service could, for example, summarize unread emails or construct a draft trip itinerary combining material from your Gmail, Docs and Google Flights. 

If a user clicks on a "Google it" feature button (the "G" icon) next to a Bard answer, Bard will check line by line "if there is content across the web to substantiate" its response, the company said in a statement, and provide that breakdown with applicable links.

Existing English language features — including image uploading and inclusion of search images in Bard responses — are expanding to more than 40 supported languages.

The intrigue: Google promised in a blog post that "your content from Gmail, Docs and Drive is not seen by human reviewers, used by Bard to show you ads or used to train the Bard model."

Tech companies have been accused of a data land grab as they grapple with how to train new and improved AI products. 

What they're saying: While walking Axios through Extensions, Jack Krawczyk, Google's product lead for Bard, said he found the new features most useful in helping him to parse back-to-school correspondence for his child and prompting him to "bring his curiosity to life" by searching in better ways.

Two types of user feedback drove today's product updates. 

"The average person around the country says, '[AI] is super cool, but I still haven't figured out how to make it useful in my life,'" Krawczyk said.

Many told Google they were concerned about chatbot "hallucinations" (making things up). 

Google's Bard launches fact-check features as AI race continues

The search engine chatbot will be able to double check its answers and plug into other Google apps.

Google

Users can now ask <a class="gtmContentClick" data-vars-link-text="Google&#x27;s Bard" data-vars-click-url="https://www.axios.com/2023/02/06/google-bard-chatgpt-rival" data-vars-content-id="b08566d0-f110-40cb-b84d-50761ffacca0" data-vars-headline="Google&#x27;s Bard launches fact-check features as AI race continues" data-vars-event-category="story" data-vars-sub-category="story" data-vars-item="in_content_link" href="https://www.axios.com/2023/02/06/google-bard-chatgpt-rival" target="_self">Google&#x27;s Bard</a> chatbot to double-check its answers and can connect it to their Google apps and services, in the first update to the product <a class="gtmContentClick" data-vars-link-text="since July" data-vars-click-url="https://bard.google.com/updates" data-vars-content-id="b08566d0-f110-40cb-b84d-50761ffacca0" data-vars-headline="Google&#x27;s Bard launches fact-check features as AI race continues" data-vars-event-category="story" data-vars-sub-category="story" data-vars-item="in_content_link" href="https://bard.google.com/updates" target="_blank">since July</a>.Why it matters: Google is racing to integrate AI into its main products, hoping to extend their dominance, after being caught out earlier this year when Microsoft integrated ChatGPT into its Bing search engine and added an AI &quot;copilot&quot; to Microsoft 365.

The organizers of a high-profile open letter last March calling for a "pause" in work on advanced artificial intelligence lost that battle, but they could be winning a longer-term fight to persuade the world to slow AI down.

The big picture: Almost exactly six months after the Future of Life Institute's letter — signed by Elon Musk, Steve Wozniak and more than 1,000 others — called for a six-month moratorium on advanced AI, the work is still charging ahead. But the ensuing massive debate deepened public unease with the technology.

The letter helped normalize expressing deep AI fears. Voters began to repeatedly express concern to pollsters, the White House mobilized tech CEOs to make safety commitments, and foreign regulators from Europe to China raced to finish AI regulations.

Between the lines: In recent months, the AI conversation around the world has intensely focused on the social, political and economic risks associated with generative AI, and voters have been vocal in telling pollsters about their AI concerns.

Driving the news:  The British government is gathering a who's who of deep thinkers on AI safety at a global summit Nov. 1-2. The event is "aimed specifically at frontier AI," U.K. Deputy Prime Minister Oliver Dowden told a conference in Washington on Thursday afternoon.

OpenAI, Meta and regulators have typically used the term "frontier AI" to distinguish the largest and potentially riskiest AI models from less capable technologies.

"You can never in one summit change the world, but you can take a step forward" and "create an institutional framework" for AI safety, Dowden said.

What they're saying: Anthony Aguirre, the executive director of the Future of Life Institute who organized the "pause" letter, told Axios in an interview that he's "pretty hopeful" the U.K. process is now the best bet for "slowing down" AI development — a subtle reframing of his original "six-month pause" goal.

Aguirre thinks it's "absolutely critical" that China play a leading role in the U.K. summit.

While acknowledging the surveillance implications of AI regulation by Beijing, Aguirre noted that China's approval process for products of foundational models such as chatbots is proof that governments can slow down the release of AI if they want to.

"Rushing, rushing, rushing to fundamental disruption is not necessarily a race you want to win," he said, adding, "The general public doesn't want runaway technologies."

Aguirre dismisses the White House-organized voluntary safety commitments as "not nearly up to the task," but is hopeful U.S. legislation will pass in 2024.

The other side: Inflection AI co-founder Reid Hoffman believes that whatever public attention the letter may have generated, the authors undermined their credibility with the AI developer community — which they will need to achieve their aims.

Hoffman told Axios the letter authors were "virtue signaling" — and claiming that only they truly cared about humanity. "It hurts the cause," Hoffman said.

Flashback: The original letter described a "dangerous race to ever-larger unpredictable black-box models" and urged AI labs to draw a line at the then recently released GPT-4, hinting that AI might one day destroy humanity.

But another camp of AI critics insisted that letter signatories were guilty of AI hype —  inflating the capabilities of the current generation of large language models.

The great AI "pause" that wasn't

No one "paused." But a lot more people are talking about AI's dangers.

The organizers of a high-profile <a class="gtmContentClick" data-vars-link-text="open letter" data-vars-click-url="https://www.axios.com/2023/03/29/elon-musk-gpt-4-chat-open-ai" data-vars-content-id="ffff940f-38c0-4803-bdd2-dab89a64645d" data-vars-headline="The great AI &quot;pause&quot; that wasn&#x27;t" data-vars-event-category="story" data-vars-sub-category="story" data-vars-item="in_content_link" href="https://www.axios.com/2023/03/29/elon-musk-gpt-4-chat-open-ai" target="_self">open letter</a> last March calling for a &quot;pause&quot; in work on advanced artificial intelligence lost that battle, but they could be winning a longer-term fight to persuade the world to slow AI down.The big picture: Almost exactly six months after the Future of Life Institute&#x27;s letter — signed by Elon Musk, Steve Wozniak and more than 1,000 others — called for a six-month moratorium on <a class="gtmContentClick" data-vars-link-text="advanced AI" data-vars-click-url="https://www.axios.com/2023/02/17/chatgpt-ai-next-platform-tech" data-vars-content-id="ffff940f-38c0-4803-bdd2-dab89a64645d" data-vars-headline="The great AI &quot;pause&quot; that wasn&#x27;t" data-vars-event-category="story" data-vars-sub-category="story" data-vars-item="in_content_link" href="https://www.axios.com/2023/02/17/chatgpt-ai-next-platform-tech" target="_self">advanced AI</a>, the work is still charging ahead. But the ensuing <a class="gtmContentClick" data-vars-link-text="massive debate" data-vars-click-url="https://www.axios.com/2023/03/30/chatgpt-ai-pause-debate-existential-risk" data-vars-content-id="ffff940f-38c0-4803-bdd2-dab89a64645d" data-vars-headline="The great AI &quot;pause&quot; that wasn&#x27;t" data-vars-event-category="story" data-vars-sub-category="story" data-vars-item="in_content_link" href="https://www.axios.com/2023/03/30/chatgpt-ai-pause-debate-existential-risk" target="_self">massive debate</a> deepened public unease with the technology.

Ryan Heath

Stories

The great AI "pause" that wasn't

Exclusive: How AI policy aligns with what voters want

UN deadlocked over regulating AI

Exclusive: First AI plan in Republican primary

Google's Bard launches fact-check features as AI race continues

Newsletter editions

AI opens Windows

Global cop no-show

Newsrooms' AI plunge

Security startup frenzy

Mental health screens